Developments and Advances in Defense and Security: Proceedings of MICRADS 2020 (Smart Innovation, Systems and Technologies, 181) 9811548749, 9789811548741
This book gathers the proceedings of the Multidisciplinary International Conference of Research Applied to Defense and S
130 91
English Pages 571 [540] Year 2020
Preface
Contents
About the Editors
Cybersecurity and Cyberdefense
Cyber Security Vulnerabilities in Colombia’s Maritime Critical Infrastructure (MCI)
1 Introduction
2 Problem Statement
3 Methodology
4 Comprehensive Analysis and Discussion of Results
5 Conclusions
References
Cyberspace and Innocent Passage: Regulations for the Security of the Coastal State
1 Introduction
2 Methodology
3 Innocent Passage Through the Territorial Sea
4 Results and Discussion
5 Conclusions
References
Security Versus Usability in E-Government: Insights from the Literature
1 Introduction
2 Background on Security and Usability in the Electronic Government
3 State-of-the-Art Systematic Review
3.1 Research Questions
3.2 Strategy and Search Strings
3.3 Selection Criteria
3.4 Selection Process
3.5 Results of the Systematic Review
4 Discussion
5 Conclusions and Future Works
References
Simulation and Computer Vision in Military Applications
Ackermann UGV with 2D Mapping for Unknown Environments
1 Introduction
2 Our Approach
2.1 Cinematic Model of Ackermann Steering
2.2 Environment Perception
2.3 Remote Control
3 Conclusions and Future Work
References
Visual-Based Real-Time Detection Using Neural Networks and Micro-UAVs for Military Operations
1 Introduction
2 Vision-Based Objective Detection System
3 Tests and Results of the Objective Detection System
3.1 Metrics Assessed on Detection
3.2 Description of the Tests Performed
4 Conclusions and Future Works
References
Visual and Inertial Data-Based Virtual Localization for Urban Combat
1 Introduction
2 Related Works
3 Our Approach
3.1 Visual-Inertial System
4 Results and Discussion
5 Conclusions and Future Work
References
Kinect and Manipulator-Based Sample Collection System for Military Robot
1 Introduction
2 Design of an Object Manipulator Robot for the Robotic Platform
3 Test and Results
4 Conclusions
References
Autonomous Navigation Based on Proportional Controller with GPS Setpoint for UAV in External Environments
1 Introduction
2 Related Works to the Subject
3 Our Flight Navigation Method
4 Test Scores and Discussion
5 Conclusions and Future Works
References
Monte Carlo-Based Localization for Kidnapped Robot Problem
1 Introduction
2 Our Approach
3 Results and Discussion
4 Conclusions and Future Work
References
Automatic Counting of People in Crowded Scenes, with Drones That Were Applied in Internal Defense Operations on October 20, 2019 in Ecuador
1 Introduction
1.1 Context
1.2 People Counting in Crowded Scenes
1.3 Drones and Crowd Counting
2 Methodology
2.1 The Application of the Gaussian Density Estimation
2.2 Classification Through the CNN
2.3 Application of the Process
2.4 Response Time Optimization
3 Results Assessment
4 Discussion
5 Conclusions and Further Work
References
Expert Nutritional System for Military Athletes Based on Fuzzy Logic and Inferential Statistics
1 Introduction
2 Related Work
3 Methodology, Experimental Configuration, and Tools (Current References)
4 Getting Data
5 Data Analysis
5.1 Choosing the Number of Main Components
5.2 Analysis for the Proposed Research
6 Results
7 Conclusions and Future Works
References
Earth Coverage Model for GPS-Like Capabilities
1 Introduction
2 Maximum Look Angle
3 Probabilistic Coverage
4 Validation
5 Results
6 Risk Analysis Application
7 Conclusion
References
Computer Networks, Mobility and Pervasive Systems
CTR Prediction for Optimizing the Negotiation of Internet Advertising Campaigns
1 Introduction
2 Methodology
2.1 Analysis, Transformation and Selection of Attributes
2.2 Campaign Segmentation
2.3 Elasticity Analysis
2.4 The Proposed Model
3 Results
3.1 Data Dimensions
3.2 Data Attributes
3.3 Evaluation of Predictive Models
3.4 CTR Prediction
4 Conclusions
References
Classification of Authors for a Recommendation Process Integrated to a Scientific Meta-Search Engine
1 Introduction
2 Development of the Method for Classifying Authors
2.1 Aspects Related to Data
2.2 Metrics and Evaluable Aspects of Authors
2.3 Classification of Authors
2.4 Proposed Algorithm for Classification of Authors
3 Experimentation and Results
4 Conclusions
References
Performance Evaluation of a Hybrid Vehicle and Sensor Network to Prevent Traffic Accidents
1 Introduction
2 Proposed Communication Protocol Between WSN and VANET
2.1 Communication Between a Static Sensor on a WSN and a Vehicle on a VANET
2.2 Vehicle-to-Vehicle Communication—Vehicles Moving in the Same Direction
2.3 Vehicle A → Vehicle B
2.4 Vehicle-to-Vehicle Communication—Vehicles Moving in Opposite Directions
3 Simulation and Results
4 Conclusions
References
Comparison of Bio-inspired Algorithms Applied to the Hospital Mortality Risk Stratification
1 Introduction
2 Methodology
2.1 Subject
2.2 Variables
2.3 Bootstrap Resampling of the Development Group
2.4 Logistic Regression Model
2.5 Artificial Neural Network Model
2.6 Artificial Neural Network Training
2.7 Model Validation
2.8 Comparison of Models
3 Results
3.1 Comparison of Models
3.2 Analysis by Diagnostic Groups
3.3 Comparison of Probabilities Between Models
4 Conclusions
References
Information and Communication Technology in Education
Articulation of Teaching and Technological Resources in the Teaching–Learning Process of the English Language
1 Introduction
2 State of the Art
3 Materials and Methods
4 Results and Discussion
5 Conclusions
References
Gravity Compensation Using Low-Cost Automation for Robot Control
1 Introduction
2 State of Technology
2.1 Kuka youBot™
2.2 Gravity Compensation
2.3 ROS
3 Case Study
4 Proposed Implementation
4.1 Gravitational Compensation
4.2 Graphical User Interface
4.3 ROS System
5 Discussion of Results
6 Conclusions and Future Work
References
Situated Learning Through the Use of Cooperative Techniques and Academic Controversy Applied to the Provision of Cryptographic Confidentiality
1 Introduction
2 Models and Learning Techniques
2.1 Situated Learning Model
2.2 Cooperative Learning Techniques
2.3 Academic Controversy Techniques
3 Computer Security, Cryptosystems, and Cryptanalysis
3.1 Computer Security
3.2 Cryptosystems
3.3 Cryptanalysis
4 The Didactic Proposal
4.1 Learning Objectives
4.2 Methodology
5 Results
6 Conclusions
References
Evolutionary Algorithm for Content-Based Image Search
1 Introduction
2 The Algorithm
2.1 The Search
2.2 Representation
2.3 Evaluation Function
2.4 Selection of Parents
2.5 Evolution Rules
2.6 A Global Approach
3 Results
4 Conclusions
References
Online Platform to Teach Aviation English at a Military School in Salinas
1 Aviation English at the Military Aviation School “ESMA”
2 Literature Review
3 Methodology
3.1 Digital Resources Selected for the Design
3.2 The Online Platform
4 Discussion
5 Conclusions
References
Manufacturing Cost Prediction Through Data Mining
1 Research Problem
2 Method and Approach
2.1 Dataset and Data Analysis
3 Results
4 Conclusions
References
Data Mining and Association Rules to Determine Twitter Trends
1 Introduction
2 Method
2.1 Pre-processing
2.2 Obtaining Association Rules
2.3 Polarization of Terms Based on Sentiments
2.4 Generalization of Rules Based on Sentiments
3 Experimentation
3.1 Data Collection
4 Results and Discussion
4.1 Donald Trump
4.2 Hillary Clinton
4.3 Generalized Interpretation Based on Sentiments
5 Conclusions
References
Jayor2: A Proposal of Information Management System for Command and Control Centers (C3i2) in the Armed Forces
1 Introduction
2 Background
3 Methodology
4 Execution
4.1 Conception and Development of the Datawarehouse
4.2 Data Mining
5 Results
5.1 Province (Geographical Location) and Incidence of Results
5.2 Military Operation and Incidence of Findings
6 Conclusions and Discussion
References
Leadership and e-Leadership
Military Training Mission in Iraq: An Exploratory Case Study Research
1 Introduction
2 Literature Review
3 Methodology
4 Findings
4.1 Intangible Actions
4.2 Tangible Actions
4.3 The House of Training
5 Concluding Remarks
References
Cultural Awareness for Civilian-Military Cooperation in Sub-Saharan Africa
1 Introduction
2 CIMIC Doctrine Due to Cultural Awareness
3 Cultural Awareness
4 Cultural Reality of Post-tribal Sub-Saharan Africa
5 CIMIC in Sub-Saharan Africa
6 Conclusions
References
Situational Awareness of Leadership in Ecuador and Its Applicability in the Multilevel Military Leadership Model
1 Introduction
1.1 Leadership in Ecuador
1.2 Leadership Styles According to the Full Range Leadership Model
1.3 Military Leadership Levels
1.4 Institutional Competencies
2 Multilevel Military Leadership
3 Materials and Methods
4 Results
5 Conclusions, Discussion, and Future Work
References
Defense Engineering
A Portable GSM Base Station Solution for Military Communications in Hostile Environments
1 Introduction
2 Methodology
2.1 Implementation
2.2 Analysis
3 Results
3.1 MOS Test
3.2 Coverage Analysis
4 Conclusion
References
Characterization and Simulation of a Propagation Model to Determine the Secondary Effects that Generate High Power and Frequency Electromagnetic Signs When Impact on Electronic Systems and Surface of Vehicles Family AMX13 of The Land Force
1 Shielding Effectiveness
1.1 Equation for Absorption Losses
1.2 Equation for Reflection Losses
1.3 Reflection Factor
2 Simulations and Results
2.1 Armored Vehicle 3D Model
2.2 Simulation of the Model
3 MATLAB Application
3.1 Design of Application
3.2 Results Using the Application
4 Conclusions and Future Investigations
References
Public Order Disruption Event Detection Based on IoT Technology. An Approach for the Improvement of Public Security Conditions
1 Introduction
2 Tendencies, Perspective and Prospective of IoT
3 Architecture, System, and Operation
3.1 Architecture
3.2 Hardware
3.3 Cloud Services
3.4 Web Client
4 Experimentation
4.1 Testing Scenario
5 Results
6 Conclusions
References
Vibration Effects of the Fixed-Wing Aircraft of the Army Aviation from Ecuador on the Human Body
1 Introduction
2 Development of the Experiment
3 Interpretation of the Results
3.1 Comfort
3.2 Health
4 Results
4.1 Cessna 206 Aircraft
4.2 Beechcraft King Air B-200 Aircraft
4.3 CN 235 Aircraft
4.4 Aravá T-201 Aircraft
4.5 CASA 212 Aircraft
5 Conclusions and Recommendations
References
Determination of Optimal Procedures for Maintenance and Repair Operations of the GDU-620 Garmin Applied to the DA20C-1 Aircraft
1 Introduction
2 Objectives
2.1 Maintenance
2.2 Manufacturer Repairs
3 Methods and Procedures
3.1 Problem Identification
3.2 Alternatives Development
3.3 Identification of Gap Alternatives and Prioritization Grid
3.4 Identification of Gap Alternatives and Prioritization Grid
3.5 Alternative Evaluation
4 Implementation
5 Results and Discussion
6 Conclusions
References
Planning, Economy and Logistics Applied to Defense
Critical Revision of the Contract Regime in the Military Service—The Case of the Portuguese Armed Forces
1 Introduction
2 Definition of Criteria for Critical Analysis of the Model of Military Service in Contract Regime in the Portuguese Armed Forces
3 Development of the Critical Analysis to the Present Model Implemented in Portuguese Armed Forces
3.1 First Criterion of Analysis: Relation Established Between Model and Environment
3.2 Second Criterion: Structuration of the Model
3.3 Third Criterion: Model Operationalization Strategy
4 Integrated Discussion on of the Model
5 Conclusions
References
Defense Organizations Budgeting and Management Control Systems in Restrictive Budgets Context—Literature Gaps
1 Introduction
2 Methodology
3 Literature Review: Preliminary Results
4 Conclusion
4.1 Limitations and Conflict of Interests
References
Coupling Architecture Between INS/GPS for Precise Navigation on Set Paths
1 Introduction
2 Methodology
2.1 Equipment Calibration and Raw Data Extraction
2.2 Inertial Measurement Unit (IMU)
2.3 Madgwick Filter Selection
2.4 The Kalman Filter
2.5 Madgwick Filter
3 Results
3.1 Validation of Results Using Statistical Methods
3.2 Experimental Method
4 Conclusions
References
Vehicle Flow Prediction Through Probabilistic Modeling
1 Introduction
1.1 Development of a VANET Using Markov Chains
1.2 PSH (Particle Smoothing Hydrodynamics) in VANET
2 Methodology
2.1 Vehicle Flow Probability Model in a VANET
3 Analysis of Results
4 Conclusions
References
Strategies and Organizational Changes for the Logistics Sustainability of Military Aircraft: Towards the Digital Transformation of In-Service Support
1 Introduction
2 Material and Method
2.1 Objectives, Methodology and Contents
2.2 The In-Service Support (ISS) Transformation Within the Integrated Logistics Support (ILS)
3 Results: The ISS Digital Transformation
4 Discussions: Strategies and Organization Changes for the Logistics Sustainability of Military Aircraft Towards the ISS Digital Transformation
5 Conclusions
References
University–Industry Collaboration Barriers: Project Management Solutions for Defense R&D—A Case Study
1 Introduction
2 Methodology
3 The Research and Development Center
4 Obstacles and Barrier to University–Industry Collaboration
4.1 Cultural Barriers
4.2 Institutional Barriers
4.3 Operational Barriers
5 Project Management Solutions for the Barriers
5.1 Cultural Barriers
5.2 Institutional Barriers
5.3 Operational Barriers
6 Conclusion
References
Business Intelligence: Use of Data Mining Techniques for the Prediction of Internment Times
1 Introduction
2 Conceptual Framework
2.1 Hospital Length of Stay
2.2 Cross-Industry Standard Process for Data Mining (CRISP-DM)
3 Method
3.1 Business Understanding
3.2 Data Understanding
3.3 Data Preparation
3.4 Modeling
3.5 Evaluation
3.6 Implementation
4 Results and Discussion
5 Conclusion
References
Mobile Military Reactors
1 Introduction
2 War-energy Data
3 The Intensifying War-Energy Nexus
4 Energy and the Risk of War
5 The Vital Role of Mobil Military Reactors
6 Conclusion
References
The Competitive Advantage of Additive Manufacturing in the Naval Sector
1 Introduction
2 Background
3 AM Processes
4 The Use of AM in the Naval Sector
5 Conclusions
References
Strategy, Geopolitics and Oceanopolitics
RETRACTED CHAPTER: Ocean Policy of the UNCLOS in Ecuador Based on New Geodynamic and Geochronological Evidences
Geopolitical Perspective on the Sea: Key Highlights from the Past and for the Future
1 Introduction
2 Sea Versus Land: Classic Geopolitical Dichotomy as Vital Element of (Classic) Geopolitical Narrative
3 (Geopolitical) Outlook for the Twenty-First Century
References
Enabling the Securitisation of the Sea Through Hybridity: EU Narrative Under Scrutiny
1 Introduction
2 Key Concepts, Theoretical Framework, and Methodology
3 EU Agenda on ‘Hybrid Threats’
4 ‘Hybrid Threats’ as an Enabler of Sea Securitisation
5 Concluding Remarks
References
Algorithms for Crime Prediction in Smart Cities Through Data Mining
1 Introduction
2 Method
2.1 Data
2.2 Regression Methods
3 Results and Discussions
4 Conclusions
References
Social Mobilizations as a Silent Instrument for Seizing Power: The Ecuadorian Case
1 Introduction
2 Analysis of Social Mobilizations
2.1 Current Situation
3 Theories, Trends and Ideologies
3.1 The Theory of Revolution and Social Emancipation for John Holloway
3.2 The Policy of Non-violent Action of Gene Elmert Sharp
3.3 The Forms of Power of Joseph Nye
4 Possible Alternative Solutions: Ecuador Case
4.1 Political Field
4.2 Legal Field
4.3 Educational Field
4.4 Field of Defence
4.5 Economic Field
5 Conclusions
References
The Armed Forces as a Immediate Response State Institution and Its Participation as an Articulator in the Risk Management in Ecuador
1 Introduction
2 Analysis of the Situation of Risks, Threats and Dangers in Ecuador
3 Intervention of the Armed Forces of Ecuador in Risk Management
3.1 Legal Framework for the Use of the Armed Forces in Support Operations for the National Risk and Emergency Management Service
3.2 The Employment of the Armed Forces of Ecuador During Emergencies Due to Disasters or Catastrophes
3.3 The Special and Technical Units of the Armed Forces of Ecuador in Support of Risk Management
4 Characteristics of the Armed Forces of Ecuador in Risk Management
5 Conclusions
References
Retraction Note to: Chapters
Retraction Note to:Chapter “Classification of Authorsfor a Recommendation Process Integratedto a Scientific Meta-Search Engine” in: Álvaro Rocha,Developments and Advances in Defense and Security,Smart Innovation, Systems and Technologies 181,
Retraction Note to:Chapter “Data Mining and Association Rulesto Determine Twitter Trends” in: Álvaro Rocha,Developments and Advances in Defense and Security,Smart Innovation, Systems and Technologies 181,
Retraction Note to:Chapter “Ocean Policy of the UNCLOS in Ecuador Basedon New Geodynamic and Geochronological Evidences”in: Álvaro Rocha, Developments and Advances in Defenseand Security, Smart Innovation, Systems andTechnologies 181,
Retraction Note to: Vehicle FlowPrediction Through ProbabilisticModeling
Author Index
Recommend Papers
![Developments and Advances in Defense and Security: Proceedings of the Multidisciplinary International Conference of Research Applied to Defense and Security ... Systems and Technologies Book 94) [1st ed. 2018]
9783319786056, 9783319786049, 3319786059](https://ebin.pub/img/200x200/developments-and-advances-in-defense-and-security-proceedings-of-the-multidisciplinary-international-conference-of-research-applied-to-defense-and-security-systems-and-technologies-book-94-1st-ed-2018-9783319786056-9783319786049-3319786059.jpg)
- Author / Uploaded
- Álvaro Rocha (editor)
- Manolo Paredes-Calderón (editor)
- Teresa Guarda (editor)
File loading please wait...
Citation preview
Smart Innovation, Systems and Technologies 181
Álvaro Rocha Manolo Paredes-Calderón Teresa Guarda Editors
Developments and Advances in Defense and Security Proceedings of MICRADS 2020
123
Smart Innovation, Systems and Technologies Volume 181
Series Editors Robert J. Howlett, Bournemouth University and KES International, Shoreham-by-sea, UK Lakhmi C. Jain, Faculty of Engineering and Information Technology, Centre for Artificial Intelligence, University of Technology Sydney, Sydney, NSW, Australia
The Smart Innovation, Systems and Technologies book series encompasses the topics of knowledge, intelligence, innovation and sustainability. The aim of the series is to make available a platform for the publication of books on all aspects of single and multi-disciplinary research on these themes in order to make the latest results available in a readily-accessible form. Volumes on interdisciplinary research combining two or more of these areas is particularly sought. The series covers systems and paradigms that employ knowledge and intelligence in a broad sense. Its scope is systems having embedded knowledge and intelligence, which may be applied to the solution of world problems in industry, the environment and the community. It also focusses on the knowledge-transfer methodologies and innovation strategies employed to make this happen effectively. The combination of intelligent systems tools and a broad range of applications introduces a need for a synergy of disciplines from science, technology, business and the humanities. The series will include conference proceedings, edited collections, monographs, handbooks, reference books, and other relevant types of book in areas of science and technology where smart systems and technologies can offer innovative solutions. High quality content is an essential feature for all book proposals accepted for the series. It is expected that editors of all accepted volumes will ensure that contributions are subjected to an appropriate level of reviewing process and adhere to KES quality principles. ** Indexing: The books of this series are submitted to ISI Proceedings, EI-Compendex, SCOPUS, Google Scholar and Springerlink **
More information about this series at https://link.springer.com/bookseries/8767
Álvaro Rocha Manolo Paredes-Calderón Teresa Guarda •
•
Editors
Developments and Advances in Defense and Security Proceedings of MICRADS 2020
123
Editors Álvaro Rocha Departamento de Engenharia Informática Universidade de Coimbra Coimbra, Portugal
Manolo Paredes-Calderón Centro de Investigación Científica y Tecnológica del Ejército (CICTE) Universidad de las Fuerzas Armadas ESPE Sangolqui, Ecuador
Teresa Guarda Universidad Estatal Península de Santa Elena La Libertad, Ecuador
ISSN 2190-3018 ISSN 2190-3026 (electronic) Smart Innovation, Systems and Technologies ISBN 978-981-15-4874-1 ISBN 978-981-15-4875-8 (eBook) https://doi.org/10.1007/978-981-15-4875-8 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore
Preface
This book contains a selection of papers accepted for presentation and discussion at the 2020 Multidisciplinary International Conference of Research Applied to Defense and Security (MICRADS’20). This conference had the support of ESPE (University of Armed Forces) of Ecuador, IME (Military Institute of Engineering) of Brazil and AISTI (Iberian Association for Information Systems and Technologies). It took place in Quito, Ecuador, during May 13–15, 2020. The 2020 Multidisciplinary International Conference of Research Applied to Defense and Security (MICRADS’20) is an international forum for researchers and practitioners to present and discuss the most recent innovations, trends, results, experiences and concerns in several perspectives of defense and security. The Program Committee of MICRADS’20 was composed of a multidisciplinary group of more than 200 experts from 40 countries around the world and those who are intimately concerned with research applied to defense and security. They have had the responsibility for evaluating, in a ‘double-blind review’ process, the papers received for each of the main themes proposed for the conference: (A) systems, communication and defense; (B) strategy and political-administrative vision in defense and (C) engineering and technologies applied to defense. MICRADS’20 received 119 contributions from 11 countries around the world. The papers accepted for presentation and discussion at the conference are published by Springer (this book) and by AISTI and will be submitted for indexing by ISI, EI-Compendex, SCOPUS and/or Google Scholar, among others. We acknowledge all of those who contributed to the staging of MICRADS’20 (authors, committees, workshop organizers and sponsors). We deeply appreciate their involvement and support that were crucial for the success of MICRADS’20. Quito, Ecuador May 2020
Álvaro Rocha Manolo Paredes-Calderón Teresa Guarda
v
Contents
Cybersecurity and Cyberdefense Cyber Security Vulnerabilities in Colombia’s Maritime Critical Infrastructure (MCI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yesid Bernardo Gomez Gamboa, Fabián Ramírez-Cabrales, and José Alejandro Machado Jiménez
3
Cyberspace and Innocent Passage: Regulations for the Security of the Coastal State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fabián Ramírez-Cabrales and José Alejandro Machado Jiménez
17
Security Versus Usability in E-Government: Insights from the Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fernando Huamán Monzón, Manuel Tupia, and Mariuxi Bruzza
29
Simulation and Computer Vision in Military Applications Ackermann UGV with 2D Mapping for Unknown Environments . . . . . Wilbert G. Aguilar, Israel Asimbaya, Pablo Albán, and Yéssica Fernández Visual-Based Real-Time Detection Using Neural Networks and Micro-UAVs for Military Operations . . . . . . . . . . . . . . . . . . . . . . . Marco Calderón, Wilbert G. Aguilar, and Darwin Merizalde Visual and Inertial Data-Based Virtual Localization for Urban Combat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wilbert G. Aguilar, Marco Calderón, Darwin Merizalde, Fabricio Amaguaña, and Jonathan Tituaña Kinect and Manipulator-Based Sample Collection System for Military Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Orlando Caiza, Wilbert G. Aguilar, Pablo Albán, and Yéssica Fernández
45
55
65
75
vii
viii
Contents
Autonomous Navigation Based on Proportional Controller with GPS Setpoint for UAV in External Environments . . . . . . . . . . . . . Darwin Merizalde, Wilbert G. Aguilar, and Marco Calderón
89
Monte Carlo-Based Localization for Kidnapped Robot Problem . . . . . . 101 Wilbert G. Aguilar, Darwin Merizalde, Marco Calderón, and Alexis Carrera Automatic Counting of People in Crowded Scenes, with Drones That Were Applied in Internal Defense Operations on October 20, 2019 in Ecuador . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Henry Cruz, Rolando P. Reyes Ch., and María Pinillos Expert Nutritional System for Military Athletes Based on Fuzzy Logic and Inferential Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Diana Vallejos, Freddy Tapia, Hernán Aules, Michelle Torres, and Cristian Bejarano Earth Coverage Model for GPS-Like Capabilities . . . . . . . . . . . . . . . . . 135 Vaughn H. Standley, Edward A. Boucheron, Robert K. Kirkwood, and Benjamin E. Norman Computer Networks, Mobility and Pervasive Systems CTR Prediction for Optimizing the Negotiation of Internet Advertising Campaigns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Jesús Silva, Jesús Vargas, Darwin Rizzo Vergara, Guillermo Araya, César Rosado, Omar Bonerge Pineda Lezama, and Benjamín Quintero RETRACTED CHAPTER: Classification of Authors for a Recommendation Process Integrated to a Scientific Meta-Search Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Amelec Viloria, Tito Crissien, Omar Bonerge Pineda Lezama, Luciana Pertuz, Nataly Orellano, and Carlos Vargas Mercado Performance Evaluation of a Hybrid Vehicle and Sensor Network to Prevent Traffic Accidents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 Jesús Silva, Noel Varela, and Omar Bonerge Pineda Lezama Comparison of Bio-inspired Algorithms Applied to the Hospital Mortality Risk Stratification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Jesús Silva, Yaneth Herazo-Beltrán, Freddy Marín-González, Noel Varela, Omar Bonerge Pineda Lezama, Pablo Palencia, and Carlos Vargas Mercado
Contents
ix
Information and Communication Technology in Education Articulation of Teaching and Technological Resources in the Teaching–Learning Process of the English Language . . . . . . . . . . 189 Albán Rocha Gina Alexandra, Salazar Mera Javier Vinicio, Collaguazo Vega Wilmer Patricio, and Paredes Viteri Santiago Efrain Gravity Compensation Using Low-Cost Automation for Robot Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Gustavo Caiza, Dalia Alvarez-Montenegro, Juan Escobar-Naranjo, Carlos A. Garcia, and Marcelo V. Garcia Situated Learning Through the Use of Cooperative Techniques and Academic Controversy Applied to the Provision of Cryptographic Confidentiality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Lucy Delgado Barra, Norka Bedregal Alpaca, Karim Guevara Puente de la Vega, and Olha Sharhorodska Evolutionary Algorithm for Content-Based Image Search . . . . . . . . . . . 229 Jesús Silva, Noel Varela, and Omar Bonerge Pineda Lezama Online Platform to Teach Aviation English at a Military School in Salinas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 Víctor Almeida, Marisol Gutiérrez, José Zambrano, and Rosalba Rodríguez Manufacturing Cost Prediction Through Data Mining . . . . . . . . . . . . . . 251 Andrea Díaz, Simón Fernández, Laura Guerra, and Eleazar Díaz RETRACTED CHAPTER: Data Mining and Association Rules to Determine Twitter Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 Jesús Silva, Jesús Vargas, Domingo Natteri, Darío Flores, Omar Bonerge Pineda Lezama, Bridy Ahumada, and Lesbia Valero Jayor2: A Proposal of Information Management System for Command and Control Centers (C3i2) in the Armed Forces . . . . . . 271 Jaime Mayorga, Robert Vargas Borbúa, Rolando P. Reyes Ch., and Tatiana Gualotuña Leadership and e-Leadership Military Training Mission in Iraq: An Exploratory Case Study Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 João Reis, Bruno Reis, Marta Nowakowska, and Aneta Kazanecka Cultural Awareness for Civilian-Military Cooperation in Sub-Saharan Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 Marta Nowakowska, João Reis, and Aneta Kazanecka
x
Contents
Situational Awareness of Leadership in Ecuador and Its Applicability in the Multilevel Military Leadership Model . . . . . . . . . . . . . . . . . . . . . 305 Celio H. Puga, Paolo Suárez, Valentina Ramos, Isabel Abad, and Alex F. Jimenez Defense Engineering A Portable GSM Base Station Solution for Military Communications in Hostile Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 Carlos Daniel Altamirano, Santiago Barragán, Felipe Grijalva, and Manolo Paredes Characterization and Simulation of a Propagation Model to Determine the Secondary Effects that Generate High Power and Frequency Electromagnetic Signs When Impact on Electronic Systems and Surface of Vehicles Family AMX13 of The Land Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 Jorge Damian Alvarez, Pablo Albán, Jonathan Lopez, and Yéssica Fernández Public Order Disruption Event Detection Based on IoT Technology. An Approach for the Improvement of Public Security Conditions . . . . . 343 Pablo Albán, López Roberto, Alexander Mejía, Isaac Vallejo, Itaty Alban, Shakira Cofre, and Esteban Molina Vibration Effects of the Fixed-Wing Aircraft of the Army Aviation from Ecuador on the Human Body . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353 Víctor Yépez, Isabel Arcentales, María León, and Pablo Caiza Determination of Optimal Procedures for Maintenance and Repair Operations of the GDU-620 Garmin Applied to the DA20C-1 Aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363 Margarita Palma, Flor Garcés, Samaniego Julio, and Pablo Albán Planning, Economy and Logistics Applied to Defense Critical Revision of the Contract Regime in the Military Service—The Case of the Portuguese Armed Forces . . . . . . . . . . . . . . . 377 Lúcio Agostinho Barreiros Santos and Maria Manuela Martins Saraiva Sarmento Coelho Defense Organizations Budgeting and Management Control Systems in Restrictive Budgets Context—Literature Gaps . . . . . . . . . . . . . . . . . . 391 Luis M. Godinho and Tiago Gonçalves Coupling Architecture Between INS/GPS for Precise Navigation on Set Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399 Jesús Silva, Noel Varela, Omar Bonerge Pineda Lezama, Hugo Hernández Palma, and Eduardo Nicolas Cueto
Contents
xi
RETRACTED CHAPTER: Vehicle Flow Prediction Through Probabilistic Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 Jesús Silva, Noel Varela, Omar Bonerge Pineda Lezama, Vladimir Álvarez, and Boris de la Hoz Strategies and Organizational Changes for the Logistics Sustainability of Military Aircraft: Towards the Digital Transformation of In-Service Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419 Manuel Antonio Fernández-Villacañas Marín University–Industry Collaboration Barriers: Project Management Solutions for Defense R&D—A Case Study . . . . . . . . . . . . . . . . . . . . . . 431 Anibal Jara-Olmedo, Mauricio Quisimalin, and Danilo Chavez Business Intelligence: Use of Data Mining Techniques for the Prediction of Internment Times . . . . . . . . . . . . . . . . . . . . . . . . . 443 Nuno M. P. Caetano and Nuno A. R. S. Loureiro Mobile Military Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461 Vaughn H. Standley, Martin D. Poon, and Joshua M. Baughman The Competitive Advantage of Additive Manufacturing in the Naval Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473 Teresa Guarda and Maria Fernanda Augusto Strategy, Geopolitics and Oceanopolitics RETRACTED CHAPTER: Ocean Policy of the UNCLOS in Ecuador Based on New Geodynamic and Geochronological Evidences . . . . . . . . . 485 Theofilos Toulkeridis, Yamirka Rojas-Agramonte, and Gustavo Paz Noboa Geopolitical Perspective on the Sea: Key Highlights from the Past and for the Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 Aneta Kazanecka, Wojciech Kazanecki, Joao Reis, and Marta Nowakowska Enabling the Securitisation of the Sea Through Hybridity: EU Narrative Under Scrutiny . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507 Ana Paula Brandão Algorithms for Crime Prediction in Smart Cities Through Data Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519 Jesús Silva, Ligia Romero, Roberto Jiménez González, Omar Larios, Fanny Barrantes, Omar Bonerge Pineda Lezama, and Alberto Manotas Social Mobilizations as a Silent Instrument for Seizing Power: The Ecuadorian Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529 Luis Recalde, Marco Antonio Criollo, and María Dolores Santos
xii
Contents
The Armed Forces as a Immediate Response State Institution and Its Participation as an Articulator in the Risk Management in Ecuador . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545 Víctor Yépez, Jorge Toledo, and Theofilos Toulkeridis Retraction Note to: Chapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Retraction Note to: Vehicle Flow Prediction Through Probabilistic Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jesús Silva, Noel Varela, Omar Bonerge Pineda Lezama, Vladimir Álvarez, and Boris de la Hoz
C1 C5
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
About the Editors
Álvaro Rocha holds the title of Honorary Professor, and holds a D.Sc. in Information Science, Ph.D. in Information Systems and Technologies, M.Sc. in Information Management, and Bachelor in Computer Science. He is a Professor of Information Systems at the University of Coimbra, researcher at CISUC (Centre for Informatics and Systems of the University of Coimbra), and a collaborator researcher at both LIACC (Laboratory of Artificial Intelligence and Computer Science) and CINTESIS (Center for Research in Health Technologies and Information Systems). His main research interests are information systems planning and management, maturity models, information systems quality, online service quality, intelligent information systems, software engineering, e-Government, e-Health, and information technology in education. He is also President of AISTI (the Iberian Association for Information Systems and Technologies), Chair of the IEEE Portugal Section Systems, Man, and Cybernetics Society Chapter, and Editor-in-Chief of both JISEM (Journal of Information Systems Engineering & Management) and RISTI (Iberian Journal of Information Systems and Technologies). Moreover, he has served as Vice-Chair of Experts for the European Commission’s Horizon 2020 program, and as an Expert at the Government of Italy’s Ministry of Education, Universities and Research, at the Government of Latvia’s Ministry of Finance, at the Government of Mexico’s National Council of Science and Technology, and at the Government of Polish’s National Science Centre. Manolo Paredes-Calderón was born in Ambato, Ecuador, in September 1978. He went to the “Eloy Alfaro” Military Academy, where he graduated in 2000 as an Army Second Lieutenant with a Bachelor in Military Sciences. Later, he obtained his Engineering degree in Electronics and Telecommunications at the Polytechnic Army School ESPE University. He obtained his Master of Science in Electronic Engineering from the Military Institute of Engineering IME in Rio de Janeiro, Brazil. His areas of expertise include digital signal processing and radio signal
xiii
xiv
About the Editors
propagation. He is currently the Director of the CICTE Military Applications Research Center and a Professor at the Department of Electrical, Electronics and Telecommunications, ESPE University of the Armed Forces. Teresa Guarda received her Ph.D. in Systems and Information Technologies from Minho University in 2015. She is currently a researcher at the ALGORITMI Research Centre, Minho University, Portugal, where she is a member of the Intelligent Data Systems (IDS) research team and the Information Systems and Technologies (IST) R&D group. She has published extensively in the following areas: pervasive systems, marketing intelligence, crowdsourcing, data mining and knowledge discovery, Internet of Things, knowledge management, and cyber-security. She has also served as a reviewer for many prominent journals, and as a Professor and researcher at various academic institutions in Portugal since 1991; currently, she is the Director of the CIST Research Center of the Santa Elena Peninsula State University (Ecuador) and a Visiting Professor at the Security and Defense Department of Universidad de las Fuerzas Armadas (Ecuador).
Cybersecurity and Cyberdefense
Cyber Security Vulnerabilities in Colombia’s Maritime Critical Infrastructure (MCI) Yesid Bernardo Gomez Gamboa, Fabián Ramírez-Cabrales , and José Alejandro Machado Jiménez
Abstract This study identifies the threats that affect the cyber security of the maritime critical infrastructure in Colombia. Specifically, four Colombian ports were analyzed. Based on a descriptive and qualitative analysis, the operational and security resources of this sample were compared with the guidelines and recommendations of the International Maritime Organization (IMO) contained in Circular 3 of 5 July 2017. The diagnosis shows negative results in terms of the current preparation of ports for possible cyber threats. It is proposed to adopt the IMO recommendations in harmony with the five strategic lines indicated by the Joint Cyber Command of the Colombian Military Forces (2017), and in compliance with the measures that regulate the navigation of ships in innocent passage through the Colombian territorial sea. Keywords Maritime cybersecurity · Innocent passage · Risks · Threats · Colombian seaports
1 Introduction Maritime navigation is an activity regulated by international law, in particular, by the right of every foreign vessel to pass through the territorial waters of a State, in time Y. B. G. Gamboa · F. Ramírez-Cabrales (B) · J. A. M. Jiménez Escuela Superior de Guerra, Carrera 11 No. 102-50 Cantón Norte, Bogotá, Colombia e-mail: [email protected] Y. B. G. Gamboa e-mail: [email protected] J. A. M. Jiménez e-mail: [email protected] Escuela Naval de Cadetes Almirante Padilla, Barrio Bosque, Isla de Manzanillo, Cartagena de Indias, Colombia Universidad Libre, Pie de la Popa Calle Real No 20-177, Cartagena de Indias, Colombia © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_1
3
4
Y. B. G. Gamboa et al.
of peace, and provided that its passage does not adversely affect peace, security, and proper order. In maritime traffic, each State has the role of a coastal State regarding its territorial sea and in accordance with international law, but also, because of the nature of the situations in matters of maritime safety that arise in such traffic, it is included the flag State or the port State. For any of these situations, the world of maritime security simply includes the identification of risks, vulnerabilities and threats, and not only for the navigation of the ships but also for the conditions of installation, maintenance and operation of the port infrastructure of a coastal State. One of the critical elements in terms of security and protection is the computer infrastructures from which the computer systems allow the operation of ships and ports, integrating such critical aspects as the ship–port interface. The access of foreign ships to the territorial waters of the coastal State demands the implementation of controls and restrictions that allow it to govern and manage everything that may affect the security of its ships, ports, population and territories. One way in which international law allows each State to design and implement these controls and restrictions is through the regulation and enforcement of the passage of ships through its waters, in a manner that responds to advances in the State of navigation and maritime safety. The IMO, through its directives and recommendations, specifies certain conditions for maritime and port activities to unify such controls and restrictions, so that activities at sea can flourish in conditions of peace, security and proper order. However, even though the IMO instruments are present, the Colombian State has yet to regulate aspects that particularly affect maritime safety when foreign ships make use of the right of passage through its territorial waters. Thereupon, as part of the research project, COLCIENCIAS—ARC No. 64905, which aims to propose standards for the regulation of the right of innocent passage of foreign ships in Colombia, the progress of a study on the aspects that should be considered in the area of maritime cyber security is presented.
2 Problem Statement The use of information and communication technologies is necessary tools for daily maritime operations. In the Colombian State’s territorial sea, controls on maritime operations have been gradually evolving, thanks to the adoption of IMO recommendations and the dynamism of port activities in recent years. However, it is necessary to observe the entire area of the Colombian territorial sea, including its air space, and the needs that arise in maritime traffic not only through the access of ships in Colombian ports but also through the passage that some foreign ships make in the territorial waters without berthing in a port. There is an increasing reliance on information and communication technologies both in the crossing and in the berthing of ships in port facilities.
Cyber Security Vulnerabilities in Colombia’s Maritime Critical …
5
The vast majority of ships and platforms that access Colombian waters use a wide range of sensors, such as meteorological (RFID),1 Wi-Fi and satellite internet sensors. These sensors communicate critical information to different systems that allow the navigation of ships, such as the Automatic Identification System (AIS), the Electronic Chart Display and Information System (ECDIS), the Global Navigation Satellite System (GNSS) and the Global Positioning System (GPS). Also, very vulnerable to any form of external interception, both ships and port facilities, are the alarms, LEDs, display and general force drivers (actuators) [1]. Technologies used in maritime activities also present increasing risks due to network interceptions and the proliferation of malware. Most of these risks have been classified as a result of cyber-attacks [2]. Studies of individual cyber-attacks have led to the proper identification of cyber risks. For example, the most emblematic recent cases have been the White Rose of Drax yacht (2013), the Ransomware-type blockade of the management and control systems of the ships of the MAERSK shipping company (2017), and the disruption of operations on oil drilling platforms (2010, 2012 and 2014), the attack on the servers of the shipping company IRISL (2011), the cargo control system of the Australian customs (2012), and the alteration of the information of the containers in the port of Antwerp in Belgium between 2011 and 2013. As a result of these studies, the IMO has set a deadline of 2021 for owners and administrations to incorporate cyber security measures on tankers subject to OCIMF2 controls. However, it is also necessary to consider that many cyber risks cannot become publicly known for justified reasons of national security and, commercially, because of speculation that may arise from the values of cybersecurity services and premiums on the marine insurance market. The identification of cybernetic risks both in ships and in port facilities has been prevented through guidelines that the IMO defined in 2002 through the regulations known as ‘ISPS Code’ in Spanish, known as the ISPS Code [3]. In the development of the ISPS and under the SOLAS Conventions, States have been regulating specific matters such as, for example, the entry of goods into air and sea terminals, which may include cyber risks. On the part of the Colombian State, the regulation of the ISPS is adopted in consideration of its commitments acquired by the SOLAS Conventions, and not in a complete and formal way, leaving to the discretion of the Government and the development of the maritime authorities the possibility of advancing regulations according to opportunity criteria for partial aspects. Thus, for example, there is a presidential decree 2155 of 2014 [4] that defines the unified standards of technology of non-intrusive inspection equipment, certainly applicable to the equipment that supports the port activities and the ships. This decree creates an administrative body with a cross-sector composition for the implementation of non-intrusive inspection 1 RFID
(Radio Frequency Identification) is a technology similar, in theory, to bar code identification, but this technology employs electromagnetic or electrostatic waves for the transmission of the signal containing the information. RFID is also known as DSRC (Dedicated Short Range Communications). 2 OCIMF is an advisory body to the IMO, constituted as an international marine forum for oil companies to promote the safe and environmentally responsible transportation of oil and oil products.
6
Y. B. G. Gamboa et al.
systems, which could establish guidelines for port operators to take more control. However, if it only regulates the conditions of unification of the inspection teams, it leaves a precedent that encourages the dispersion of regulations, in order to diversify the subjects, and not to concentrate those elements that are integrated into the intrusive cyber risk, especially necessary in the ports of the States as in the case of Colombia. It is necessary not to forget that these risks make the maritime ports a critical focus of vulnerabilities not only of cyber-attacks but also of spontaneous or not allowed intrusions, which can affect both the security conditions in the port activities as well as the access to the infrastructure interconnected with the ports. There is also a particular need for State regulation to contain correction criteria that are more appropriate to the needs of maritime cybersecurity. These criteria are the result of relations between owners, shipping agents, ship managers and providers for cybersecurity risk management. In the event that the regulation of States is dispersed, the IMO will issue recommendations in 2016 in Circular 1525 of 1 June 2016 [5]. Within the organization of the Colombian maritime public administration, and the powers that its bodies legally have, Circular 1525 specifies some of the following recommendations, which can be included in the Colombian regulations for the management of maritime cyber risk: 1. Specify the powers to be exercised by officials, in particular regarding inspection, testing of ship and port facility security measures and procedures, together with the possible application of enforcement measures to correct instances of noncompliance. 2. In maritime security measures for cybercrime risk, differentiate matters to be regulated by so-called ‘primary regulation’ (i.e. laws and decrees) from what is called ‘secondary regulation’ (regulations, instructions, manuals, adopted in pursuance of powers granted by primary regulation). 3. Distinguish between provisions that are mandatory for taking maritime protection measures and those which are merely indicative or optional for certain needs of optimization of the resources employed. 4. Determine the conditions for issuing Declarations of Security when they are required for the assessment of cyber risk in a ship/port interface operation or ship-to-ship activity that may affect people, property or the environment. 5. Determine the procedures to be followed for the identification and reporting of incidents, with a classification appropriate to the cyber risk, while allowing for prompt notification to local law enforcement agencies, when the ship is in Colombian territorial waters. Since there is a pressing need to identify such risks in advance, the IMO subsequently issued Circular 3 of 5 July 2017 [6], providing some recommendations that serve as guidelines for the management of maritime cyber risks, especially with regard to the controls that can be implemented by governments that are both in the position of flag State and coastal State. Afterwards, the IMO Maritime Safety Committee developed, on 11 March 2019, the Maritime Industry Guidelines on Shipboard Cybersecurity in version 3 (MSC 101/4/1) [7]. These guidelines are considered necessary for port structures because
Cyber Security Vulnerabilities in Colombia’s Maritime Critical …
7
Fig. 1 Source prepared by the authors, based in [8]
between ships and ports, there is a principle established since the ISPS Code (ISPS) making parity between the structural elements of risk management of a ship with the elements contained in the port, or vice versa, mandatory. Because of this principle, the guidelines on cyber security on board ships apply mutatis mutandis to ports. If a State imposes certain cybersecurity conditions as a requirement for flagging in a ship, these conditions will eventually be replicated in ports which the coastal State must necessarily provide with the same conditions for the ships it expects to receive in its ports. In summary, the vulnerability from cyber-attacks is concentrated in the so-called maritime critical infrastructure (MCI): port terminals, ships and their cargo, and maritime control stations. The various infrastructures that are subject to the greatest protection and defense care have also been called cyber critical infrastructures or digital critical infrastructures (Fig. 1) [8]. From the controls that can be carried out by the maritime administrations of the governments, the infrastructures can be supported by two classes of technologies: information technologies (IT) and operation technologies (OT). The above distinction makes it possible to specify the conditions from which the type of threats and vulnerabilities to MSI must be identified. Thus, for example, while the former (IT) supports information systems for port operation that could generate errors that delay or paralyze traffic, the latter could support automatic controls of gantry cranes with the risk of generating accidents in the movement of cargo on the ship or within the port. A review of all the rules of domestic law in Colombia shows that there is currently no regulation that precisely allows for greater control of maritime operations in its jurisdictional waters in the face of cybernetic risks in its port facilities. This study specifies how the vulnerabilities in Colombian port infrastructure are currently dealt with in the face of possible cyber-attacks, with the aim of creating a regulation for maritime activities that is more in line with the needs that arise from the use of
8
Y. B. G. Gamboa et al.
ICTs, so that the exchange of information can flow and so that the implementation of protection measures against this type of attack is in line with the good maritime practices encouraged by the IMO.
3 Methodology In order to identify the vulnerabilities of Colombian ports, the main ports were sampled, and in order of importance those located in the cities of Cartagena, Buenaventura, Barranquilla and Santa Marta. In each of the ports, the equipment directly related to ICTs was identified, especially the control and cargo equipment of each of the corporations that own the port operation. In the study, each port is evaluated according to its greater or lower capacity to receive container ships and the flow of its logistics operation in relation to the security systems that each port has. The equipment of the port terminals in the sample registers common characteristics even though they were built by different companies and operated by different corporations. These common structural characteristics facilitated an analysis of their vulnerabilities to cyber-attacks. Indeed, the characteristics of the equipment directly related to ICTs have an architecture based mainly on PLC (Programmable Logic Controller) devices, data packet transport protocols such as TCP (Transmission Control Protocol) and network protocols such as IP (Internet Protocol). These combinations make the systems susceptible to affectation through cybernetic attacks (Fig. 2). The analysis of vulnerabilities in the port facilities of the main ports included in the sample is presented specifying the equipment with its characteristics obtained through a survey applied to the respective port operators (Table 1). The survey applied in 2019 revealed the results (Tables 2, 3, 4 and 5).
Fig. 2 Source prepared by the authors, based in [6]
Cyber Security Vulnerabilities in Colombia’s Maritime Critical …
9
Table 1 Sample analysis Littoral
City
Responsible port operation company
Acronym
Caribbean Sea
Cartagena de Indias
Sociedad portuaria regional de Cartagena
SPRC
Pacific Ocean
Buenaventura
Sociedad portuaria de Buenaventura
SPRBUN
Caribbean Sea
Barranquilla
Sociedad portuaria regional de Barranquilla
SPRB
Caribbean Sea
Santa Marta
Sociedad portuaria de Santa Marta
SPSM
Table 2 Sociedad portuaria regional de Cartagena SPRC Equipment
Availability (Yes/No)
Total quantity and available capacity
Harbour crane
Yes
4 (50–70 ton)
STS gantry cranes with 22-container span and twinlift capacity for simultaneous unloading of two 20 ft. containers
Gantry crane
Yes
24 (6 stack up)
STS gantry cranes with 22-container span and twinlift capacity for simultaneous unloading of two 20 ft. containers
Mobile cranes
Yes
2 (100 ton)
100 ton capacity mobile cranes
Reach stacker
Yes
14 (45)
RoRo tug with tráiler
N/A
Grain elevator with packing machine
N/A
Trans strainer crane
Yes
21 (40 ton)
Forklift
Yes
20 (2, 5–7 ton)
Table 3 Sociedad portuaria regional de Buenaventura SPRBUN Equipment
Availability (Yes/No)
Total quantity and available capacity
Harbour crane
Yes
10 PP + 2 SPP
In use
Gantry crane
Yes
6 (40–62 MT)
Post-Panamaz shore rails
Mobile cranes
Yes
3-PP + 1 SPP
Multipurpose mobile crane
Reach stacker
Yes
16 (40 and 30 MT)
In use
RoRo tug with tráiler
No
N/A
N/A
Grain elevator with packing machine
Yes
Siwertel: 700 ton/h Vigan: 300 ton/h Buhler: 300 ton/h
Suction discharge
Trans strainer crane
Yes
22 (60 MT)
To deliver and receive containers, tyre crane
Forklift
Yes
10
16 and 22 ton forklift
10
Y. B. G. Gamboa et al.
Table 4 Sociedad portuaria de Barranquilla SPRB Equipment
Availability (Yes/No)
Total quantity and available capacity
Harbour crane
Yes
2 (100 ton)
Gantry crane
Yes
3–30 ton
Mobile cranes
Yes
124 ton
2 Liebherr LHM 420
Reach stacker
Yes
7–18 ton
17 reach stacker and 3 straddle carriers container handling
RoRo tug master (W/trailer)
No
N/A
N/A
Grain elevator with packing machine
Yes
280 ton/h
1
Trans strainer crane
No
N/A
N/A
Forklift
Yes
Up to 45 ton
13
Table 5 Sociedad portuaria de Santa Marta SPSM Equipment
Availability (Yes/No)
Total quantity and available capacity
Harbour crane
Yes
2 (3–15 MT)
Electrical
Gantry crane for containers
Yes
4
RTG 6 + 1
Mobile cranes
Yes
150 MT
N/A
Reach stacker
Yes
45 MT
N/A
RoRo tug master (W/trailer)
N/A
N/A
N/A
Lifting crane w/bagging machines
Yes
1 (180 ton/h) 1 (450 ton/h)
Suction equipment
Grúa trans strainer
No
N/A
N/A
Forklift
Yes
29
6 electrical
Regarding the systems for the management and control of cargo and for security in the port, the survey also provided the results (Table 6). On the other hand, the Colombian Maritime Authority exercises traffic control through the ports by means of a system that allows the coordination of reaction operations in the event of situations that may affect the security of the ports, and under the action of the units of the National Navy Coast Guard. The system is known as SICTVM (Integrated System for Traffic Control and Maritime Surveillance), which interconnects sensors, servers and administration consoles [9]. The architecture of this system employs RADAR, ECDIS, GPS and AIS informing the servers and computers used for administration, and communicating through
Cyber Security Vulnerabilities in Colombia’s Maritime Critical …
11
Table 6 Management and control of cargo at the port versus existing security technological tools at the port Systems for management and control of cargo at the port
Technological implements of security at the port
SPRC
It has the capacity to simultaneously receive up to five Neo-panamax type vessels because it has the latest versions of systematized technology such as an information management system for port management which, together with the equipment described in Fig. 1, allows a constant flow in the logistics operation chain
It has state-of-the-art technology to protect and safeguard cargo, motor vessels, terminals, operation equipment and users, through information systems that alert to potential risks that arise in the development of port operations. The following security controls are in place: 1. Access control through intelligent identification credentials 2. Closed-circuit television (CCTV) 3. High-tech scanners to verify the content of containers entering or leaving the port
SPRBUN
Logistical capacity relies on seasonal restrictions due to climatic conditions, especially rainfall, occasional sediment from river mouths in the bay, and high traffic by artisanal fishers. It also depends on the timely arrival of cargo collection vehicles, which make land trips from the centre of the country. The teams mentioned in the figure carry out the maritime operations inside the port facility
There is a coordination of the security teams employed with a high standard, through an integrated electronic security system, which allows to provide the shipowners, logistic agencies and workers’ personnel, security during the development of any procedure inside the port. The following systems are available and regularly used: 1. Biometric readers for access control to the facilities 2. Full and half body swivel mounts 3. Vehicle access points, all controlled by a server containing detailed data on the port population 4. Closed-circuit TV 5. Special system to monitor the internal navigation channel under any visibility condition, which includes an average recording time of 75 days
SPRB
It is equipped with an information management system for port management connected to the equipment mentioned in the figure
The latest technology systems used in the port include non-intrusive X-ray inspection equipment, which allows the detection of explosive and narcotic chemicals on board containers entering or leaving the port terminal
SPSM
The equipment indicated in the figure has been the result of recent modernization and technification processes of the port
The logistics operation chains are protected by the following elements: Human heat detection cameras Closed-circuit TV Non-intrusive X-ray inspection equipment, which allows the detection of chemical, explosive and narcotic substances on board containers entering and/or leaving port facilities
12
Y. B. G. Gamboa et al.
commercial protocols, due to this they can be affected through cyberspace, for example, with a cyber-attack that affects the control of ship traffic, completely paralyzing the port, an access channel or trigger discharges of pollutants into the marine environment. The towers of the maritime traffic control stations located in each port regulate the traffic in and out of national ports. The towers transmit information to the National Navy units that carry out operations to ensure the safety and security of ships entering and/or leaving the port. In the event of a cyber-attack, errors would result in accidents within the port access channels. With the intrusion, the manipulation and/or degradation of the security systems controlled by the Coast Guard would facilitate criminal activities such as drug trafficking, smuggling and human trafficking. It could also affect the automatic control of weapons on board the National Navy’s units afloat [9]. Because of existing vulnerabilities, the Joint Cybernetic Command of the General Command of the Colombian Military Forces adopted a National Protection and Defense Plan for Colombia’s Critical Cybernetic Infrastructure [10]. The Plan aims to develop an information resilience capacity through five strategic lines, tending to strengthen the cyber security of MSIs throughout the country: Identify, Protect, Detect, Respond and Recover.3 The Plan includes the implementation of guidelines, certification and awareness programs, the strengthening of response capacities and the recovery from threats directed from the National Protection Centre. However, although the Plan does not contradict the proposals of the guide on cybersecurity on board ships, the guide is more specific, especially in what it establishes as strategic circular phases of cyber risk management: identify threats, identify vulnerabilities, assess risk exposure, develop protection and detection measures, establish contingency plans, respond to recover from cyber security incidents [11].
4 Comprehensive Analysis and Discussion of Results So far, there are no guidelines or investment plans for equipment in Colombian port terminals that would make it possible to sufficiently deal with the risks arising from cyber-attacks and with the clear purpose of ensuring and generating protection and resilience for critical infrastructure. In fact, although the main ports have the latest technological versions of security tools, these only exist to address the need 3 The
aspects contained in the proposal, seek to be part of the aspects to require vessels to make use of the right of innocent passage in Colombian territorial waters, in a state of normal navigation. Its scope is given for the critical infrastructure of ships with components of OT/IT or hybrids and are based on the guidelines of the National Center for Critical Infrastructure Protection of Spain CNPIC, the twenty critical security controls SANS and the five functions of the NIST framework “Improving Critical Infrastructure Cybersecurity”, which are Identify, Protect, Detect, Respond and Recover, which is also adopted by the Joint Cybernetic Command of the General Command of the Military Forces of Colombia, in the first version of the publication of the National Plan of Protection and Defense for the Critical Cybernetic Infrastructure of Colombia.
Cyber Security Vulnerabilities in Colombia’s Maritime Critical …
13
to prevent unauthorized physical intrusions into facilities and ships, by people or goods, but there are no equipment or protocols for identifying virtual or computerized intrusions. Observing the assessment conducted in 2017 by the London company IHS Fairplay4 through a survey in which 284 people from the maritime sector participated, such measurement showed that the cyber-attacks originated both on board the ships and in the port facilities due to the lack of training and awareness of the crew, as well as a culture that tolerates the indiscriminate use of personal electronic devices at work by connecting the ship’s systems or the port terminal facility, to upload or download information or to take power for their batteries, activities that could release malware. Such practices are still observed among personnel working in port facilities. Furthermore, attention must be drawn to the fact that in Colombia there are port operators, who are mostly companies that generate profits for the markets and private capital. From this fact, it can be seen that for port operators there are no corporate and organizational guidelines to resolve situations arising from the lack of exchange of information on vulnerabilities among port operators in the face of cyber-attacks, as an act of necessary solidarity, because they contradict the business reserve policies of each of these operators: this means that companies cannot manage to protect themselves proactively because by using the software in common use throughout the world, it facilitates that hackers improve their sophistication skills and reuse more often the methods of attacking the various port operators because, by behaving in isolation from each other’s risks and threats, they do not carry out the appropriate prior warnings to adopt preventive measures [11].
5 Conclusions Based on the five main ports studied and referenced at the national level, it was possible to determine that the use of IT/OT systems for loading and unloading operations may be subject to a cyber-attack in terms of operating and cargo control equipment. Therefore, the concept of cybersecurity within the framework of the relationship between maritime authorities and stakeholders in the management and control of national maritime security is a point to be included within the regulation and control of maritime activity. E The Colombian maritime sector has not fully implemented the recommendations and guidelines of the IMO to reduce maritime cyber risk, as it is not aware of any type of Plan, program and/or effort that the Colombian Maritime Authority has managed to create guidelines to strengthen a culture of cyber risk awareness in the critical infrastructure of the maritime sector. There is no known IMO orientation, 4A
company that owns the largest maritime databases in the world, developed from the Lloyd’s Register of Ships, which has been published continuously since 1764. It has extensive databases of ships, ship movement, casualties, ownership, ports and news, as well as research and consulting services.
14
Y. B. G. Gamboa et al.
recommendation or guideline implemented by the Colombian Maritime Authority for the management of maritime cybernetic risk caused by foreign flag vessels that exercise the right of ‘Innocent Passage’ in the Colombian Territorial Sea and that may affect peace, proper order and State security. The resilience of MSIs will be able to be limited not only to the five strategic lines set out by the CCOC, but also to the best conditions for maritime risk management, guided by more specific strategic phases such as those set out in the Guide for Ships on Board applied mutatis mutandis to port structures, which are as follows: identify attacks, identify vulnerabilities, assess risk exposure, develop protection and detection measures, establish contingency plans, response and recovery from cyber security incidents. These phases also include the implementation of guidelines, certification and awareness-raising programs; the strengthening of response and recovery capacities in the face of threats through the creation of the National Protection Centre, aspects that are contained in the CCOC’s recommendations; Of course, it is necessary to generate regulations that allow the management of digital security in the threats that can be generated from ships that make use of the right of innocent passage from the Colombian territorial sea. The management of cybernetic risks for the Colombian maritime sector constitutes an approximation and reference parameter to build the agenda of maritime cyber security in terms of the exercise of the right of innocent passage of foreign ships through Colombian territorial waters and their interactions with port facilities and terminals.
References 1. BIMCO, CLIA, ICS, INTERCARGO, INTERMANAGER, INTERTANKO, IUMI, OCIMF and World Shipping Council: The Guidelines on Cyber Security Onboard Ships, Version 3. Homepage: https://www.bimco.org/products/publications/free/cyber-security (2018). último acceso 2019/12/11 2. Sen, R.: Cyber and information threats to seaports and ships. Marit. Secur., 281–302 (2016). https://doi.org/10.1016/b978-0-12-803672-3.00009-1 3. Organización Marítima Internacional: Código internacional para la protección de los buques y de las instalaciones portuarias y enmiendas de 2002 al Convenio SOLAS. Homepage: http:// www.imo.org (2003). último acceso 2019/12/11 4. República de Colombia: Ministerio de Hacienda y Crédito Público, Decreto 2155 de 2014 (2014) 5. Organización Marítima Internacional: Circular 1525. OMI, London. Homepage: http://www. imo.org (2016). último acceso 2019/12/11 6. Organización Marítima Internacional: Guidelines on Maritime Cyber Risk Managment. MscFal 1/Circ.3. OMI, London. Homepage: http://www.imo.org (2017). último acceso 2019/12/11 7. Organización Marítima Internacional: Directrices al sector marítimo sobre ciberseguridad a bordo de los buques vrs. No. 3. OMI, London. Homepage: http://www.imo.org (2019). último acceso 2019/12/11 8. Das, S.K., Krishna, K., Zhang, N.: Handbook on Securing Cyber-Physical Critical Infrastructure. In: Xue, M., Roy, S., Wan, Y., Das, S.K. (eds.) Morgan Kaufmann (2012). https://doi.org/ 10.1016/b978-0-12-415815-3.00037-6 9. Luque, A., Aponte, J.: Análisis de Riesgos y Amenazas Cibernéticas a las Líneas de Comunicaciones Marítimas Nacionales. Cartagena (2019)
Cyber Security Vulnerabilities in Colombia’s Maritime Critical …
15
10. Comando Conjunto Cibernético CCOC: Plan Nacional de Protección y Defensa para la Infraestructura Crítica Cibernética de Colombia. Imprenta de las FFMM, Bogotá, Colombia (2017) 11. Reiskind, A.: Seguridad Cibernética y Transporte Comercial Marítimo: ¿Todo está en Orden? Global Technologies for Defense and Security. Asociación Canadiense para la OTAN (NAOC) (2018)
Cyberspace and Innocent Passage: Regulations for the Security of the Coastal State Fabián Ramírez-Cabrales
and José Alejandro Machado Jiménez
Abstract This research identifies the actors, interests, and situations involved in the development of cybernetic operations that can affect the peace, good order, and security of the coastal State in its territorial sea. Particularly, special emphasis is placed on the right of innocent passage of foreign ships navigating in such space. Each of the activities of this legal concept is studied in order to determine whether or not the passage complies with the condition of innocence and its close relationship with the use of cyberspace for such purposes. Fifteen interests and their implications for security are recorded and analyzed from the perspective of cyberspace. The study concludes that in the absence of an international regime for the governance of cyberspace and the legal obsolescence of customary international law to regulate activities related to innocent passage, it recommends that States should empower themselves with laws and regulations to ensure their security in cyberspace. Keywords Cyberspace · Innocent passage · Ships · Security · State · Cyber-attack
1 Introduction National Defense is a responsibility of the State to guarantee national independence and the stability of institutions and its main purpose is to preserve sovereignty. Currently, this constitutional mandate has evolved towards the incorporation of new concepts in the field of security. New information and communication technologies have generated a challenge to national security [1]. These challenges are associated with digital convergence. Digital convergence is the result of a series of innovations F. Ramírez-Cabrales (B) · J. A. Machado Jiménez Armada de Colombia, Carrera 10 #26-27 Piso 4, Bogotá D.C., Colombia e-mail: [email protected] J. A. Machado Jiménez e-mail: [email protected] Universidad Libre, Pie de la Popa Calle Real No 20-177, Cartagena de Indias, Colombia © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_2
17
18
F. Ramírez-Cabrales and J. A. Machado Jiménez
that generate an increasingly interconnected world of knowledge, documents, data, and information converging in real-time. Activities such as international trade and finance, social networking, and egovernment are key elements for the operation of critical infrastructure services. The uses of these have raised a range of problems in terms of the security and defense of States [2]. It can affect international relations; it can be a tool or means for the achievement of national interests or a resource for peace or war. In this regard, States have found in cybernetic operations a means or tool to influence other States, in a disruptive manner, or through degradation or espionage through cyberspace [3]. Cyberspace is a concept under construction that dates back to the last century. Nowadays, several definitions can be identified. Cyberspace is an amorphous space that does not occupy a certain physical or geographical location [4]. The vast majority of definitions circulating about cyberspace agree that the core of cyberspace is composed of globally interconnected hardware, software, and data networks along with human interaction with those networks [5]. The importance of this space lies in the processing of information. In the so-called cybersociety [6], the assumption is that information by itself has a value capable of generating power (political, economic, social, etc.) and appears as a strategic asset of countries [7]. Consequently, threat, attack, and cyber conflict are becoming consolidated as alternatives to the traditional modalities of the use of force and armed conflict. “All nations have vulnerabilities that can be exploited in and through cyberspace” [8]. In international shipping, computer and operational technology on board ships are increasingly networked, and more often connected to the Internet. The increasing complexity of ships and their connection to services provided from land-based networks via the Internet means that on-board systems are increasingly exposed to cyber-attacks. In this perspective, these systems can be vulnerable by being used as a means to launch a cyber-attack, or be affected by a cyber-attack that is carried out. This represents a risk to the safety of vessels and the coastal State in the territorial sea. In particular, the greatest vulnerability is identified with unauthorized access and malicious attacks on ships’ systems and networks. According to the International Maritime Organization, the security, environmental and trade consequences of not being prepared for a cyber-incident can be considerable [9]. In the response to this threat, the IMO, the United Nations specialized agency responsible for the safety and security of navigation and the prevention of marine pollution from ships, considered that measures should be developed to enhance maritime security. Resolution MSC.428(98): “Management of Maritime Cyber-Risk in Safety Management Systems” States that approved safety management systems (SMS) should take into consideration the management of cyber-risk in accordance with the objectives and functional requirements of the International Safety Management (Code ISM). MSC-FAL.1/Circ.3: “Guidelines on maritime cyber-risk management” provides high-level recommendations on maritime cyber-risk management to safeguard maritime transport from current and emerging cyber-threats and vulnerabilities. The Industry guidelines provide instructions on how to comply with resolution MSC.428(98) at the first annual verification of the shipping company’s compliance document after January 1, 2021. The objective of maritime cyber-risk
Cyberspace and Innocent Passage: Regulations for the Security …
19
management is to contribute to the safety and security of maritime transport which is operationally resilient to cyber-risks [10, 11]. This approach allows the identification of cyber-threats and vulnerabilities on board ships. The IMO does not deal with or address the regulation of cyberspace. This ratifies that international society and States, as those responsible for its legal regulation, have not fully appreciated its nature, nor its effects on other physical spaces [12], nor its delimitation, nor its legal regime [13]. This poses a challenge to the legal regime of the territorial sea. While customary international law recognizes sovereignty and jurisdiction over the territorial sea and the area of airspace lying on the surface of that sea within 12 miles of the baselines of the continental or island territory, it has not incorporated the dimension of cyberspace and its effects on the security of the coastal State. This legal gap creates the need to identify the actors, interests, and situations involved in the development of cybernetic operations that can affect the peace and good order of the coastal State. Specifically, with regard to the right of innocent passage of foreign ships through the territorial sea. This is in order to raise the need to condition the use of cyberspace to the purposes and aims of the State that allow to guarantee sovereignty, security, and its maritime protection. To this end, this paper, a product of the research project MINCIENCIAS #64905 for the execution of R + D + i projects of the Colombian Navy, entitled “Regulation of the right to the innocent passage of foreign ships through the Colombian territorial sea: a normative proposal”, aims to identify the threats that cyberspace poses to the peace and security of the coastal State during the exercise of its duties. The Committee is concerned about the lack of transparency of innocent passage in the territorial sea and the need to include cyberspace in the regulation of this passage. Within that framework, the paper has been divided into five parts. First, the introductory aspects related to the theoretical and objective framework of the research. The second part describes the methodology used and the phases developed within the framework of the research project. The third part explains the content and scope of the figure of the right of innocent passage, highlighting the normative aspects that regulate this right and the situations that can transcend into cyberspace. The fourth part identifies the types of cybernetic operations. In this section, special emphasis is placed on the development of cybernetic operations linked to the military and merchant navy field. It highlights the threats that can originate in cyberspace, the risks and actors that can motivate actions to the detriment of the security of the coastal State in its territorial sea, motivating its discussion. Finally, the last part highlights the conclusions related to cyberspace and its impact on the maintenance of peace and security of the coastal State in the territorial sea. In this last section, final considerations will be raised without being strictly conclusive given the complexity and technological progress of the means used in cyberspace.
20
F. Ramírez-Cabrales and J. A. Machado Jiménez
2 Methodology It was considered necessary for this study to first analyze the concepts, scope, and definitions of cyberspace and the right to the innocent passage of foreign ships in the territorial sea. Once the theoretical framework of this proposal had been identified, each of the situations in which the passage of a foreign vessel would be considered to be detrimental to the peace, good order, or security of the coastal State was analyzed. To this end, 15 interests and the actors that motivate a behavior or action that according to the objective can be considered an undesirable act for the security of the State in its territorial sea were identified. Specifically, this phase analyses the connection between the security of international maritime traffic and the use of cyberspace for civil or military purposes. The results obtained in this phase allowed partial conclusions to be drawn because of the constant evolution of this complex problem associated with the use of cyberspace and its relationship with the territorial sea to regulate the activity of innocent passage by foreign ships.
3 Innocent Passage Through the Territorial Sea The increase in international maritime trade by sea and the digital economy, as well as the projection of new routes for international navigation, require the strengthening and modernization of national legislation to regulate the right of innocent passage of foreign vessels through the territorial sea, including island territories. The right of innocent passage may be exercised on condition that it does not disturb the peace, order or security of the State. Innocent Passage means the act of sailing through the territorial sea for the purpose of crossing that sea without entering the internal waters and without making a stopover or going into or out of internal waters, or calling at a port facility, or leaving it. The passage of a foreign vessel shall be deemed to be prejudicial to the peace, good order, or security of the State if that vessel infringes the sovereignty, territorial integrity, or political independence of the State. In this regard, customary international law compiles a list of activities that are deemed to be prejudicial to the peace, good order or security of the coastal State and that result in the passage of a foreign vessel being not innocent. Under customary international law, the passage of a foreign vessel shall be deemed to be prejudicial to the peace, good order or security of the coastal State if that vessel engages in any of the following activities in the territorial sea: (a) Any threat or use of force against the sovereignty, territorial integrity or political independence of the coastal State or in any other manner violating the principles of international law embodied in the Charter of the United Nations; (b) Any exercise or practice with weapons of any kind; (c) Any act aimed at obtaining information to the detriment of the defense or security of the coastal State; (d) Any act of propaganda aimed at undermining the defense or security of the coastal State; (e) The launching, receiving
Cyberspace and Innocent Passage: Regulations for the Security …
21
or loading of aircraft; (f) The launching, receiving or loading of military devices; (g) The embarkation or disembarkation of any product, currency or person in contravention of the customs, tax, immigration or sanitary laws and regulations of the coastal State; (h) Any act of intentional and serious pollution contrary to this Convention; (i) Any fishing activities; (j) The conduct of research or hydrographic surveys; (k) Any act aimed at disrupting the communication systems or any other services or facilities of the coastal State; (l) Any other activity not directly related to the passage. This diversity of activities can guide the State’s action in its national interest to reduce the margin of uncertainty in estimating the harmlessness or otherwise of the passage. Furthermore, Hakapää and Molenaar [14] argue that the list is not absolute because of paragraph (l) which refers to any other activities not related to innocent passage. Therefore, the coastal State may adopt, in accordance with customary international law, laws and regulations concerning innocent passage through the territorial sea on some or all of the following matters: (a) The safety of navigation and regulation of maritime traffic; (b) The protection of aids to navigation and other services and facilities; (c) The protection of cables and pipelines; (d) The conservation of the living resources of the sea; (e) The prevention of violations of its fisheries laws and regulations; (f) The preservation of its environment and the prevention, reduction and control of pollution thereof; (g) Marine scientific research and hydrographic surveys; (h) The prevention of violations of its customs, immigration, and sanitary laws and regulations. Currently, each of these activities can be said to be embedded in a cyberenvironment that links States, organizations, users, networks, devices, all types of software, processes, stored or circulating information, applications, services and systems that are directly or indirectly connected to networks [15]. According to Bejarano [16], the actors involved (including States, businesses, organizations, groups or individuals) will compete to control it. This inevitably leads to conflicts in cyberspace. Cyber conflict is a confrontation between two or more parties, and at least one of these parties will resort to cyber-attacks to attack the other. The context will differ from the nature and objectives of the actors. Therefore, the territorial sea and the regulation of the innocent passage of ships do not escape the threat of cyberspace as a means of affecting the peace and security of the coastal State.
4 Results and Discussion Cybernetic operations can be classified as offensive or defensive. They can be carried out through two mechanisms: physical ones, such as computers, modems or cables, and those mechanisms that only work in cyberspace, such as computer programs or viruses [17]. At the military cybernetic operations level, two challenges arise: sovereignty/jurisdiction, attribution/responsibility [18]. In the area of sovereignty, the principle of the sovereign equality of States is ratified. However, sovereignty is intimately linked to the existence of physical space, which is why its application in
22
F. Ramírez-Cabrales and J. A. Machado Jiménez
cyberspace is not so clear. It is therefore impossible for it to be appropriated and for a State to exercise sovereignty [4–19]. However, at the cybernetic infrastructures level and on any activity related to these infrastructures, the State exercises sovereignty since they are within its territory [5]. In terms of jurisdiction, the State exercises jurisdiction over activities taking place in cyberspace, such as the commission of cybercrimes. However, States do not have sufficient capacity to deter, prevent or detect unwanted activity occurring on their networks. This makes it difficult to apply the principle of State responsibility [20]. One of the events observed is the case of ransomware NotPetya, which although it started as a specific attack on Ukraine, managed to expand rapidly to systems throughout Europe and other sectors of the world. An example of its consequences was the collateral damage to the transport company Maersk. This large-scale cyberattack was attributed by Britain to Russia, in an offensive to affect Ukraine. Despite this complaint being filed, there is no public evidence of Russia’s participation in this attack. On this matter, the Tallinn Manual provides: “A State shall not permit the cybernetic infrastructure located in its territory or under exclusive government control to be used to carry out acts that illegitimately affect the rights of other States” [22]. Therefore, this manual suggests that, despite the difficulties, States have an obligation to take all appropriate measures to prevent the rights of third States from being affected. This applies to the exercise of sovereignty of the coastal State over the territorial sea and the need to ensure that the passage of foreign vessels over this space is not altered by acts affecting the rights of the flag State. In this study, 15 interests were identified, described in Table 1, which may cause conflict situations in the development of cybernetic operations harmful to the peace and security of the Coastal State during the passage of foreign ships through the territorial sea. It should be noted that the main actor, although not the only one, is the State as the guarantor of national security and its interests in cyberspace. Each of these describes at least one situation whose aim is to affect the peace and security of the coastal State. It is clear that in matters of sovereignty, States cannot dominate cyberspace, but rather share it. According to Mr. Vice Admiral Fernandez (2013) [23], there can be no such thing as Cyberspace Domain as there is no such thing as Sea Domain. In cyberspace, human interaction uses networks to obtain superiority in information by denying it to the adversary. It was also identified that within these networks, cybernetic networks for defense subsist. The purpose of these networks is to analyze the characteristics of enemy viruses and other types of possible attacks in an attempt to eliminate them. As for weapons in cyberspace, viruses or malware, they do not produce destruction but rather interruption of functions. They are therefore considered to be weapons of mass disruption. In communications, naval operations are prone to cyber-attacks. In particular, command and control and communications systems at the strategic and operational levels are vulnerable to cyber-attack. Networked cyber operations, a concept
Cyberspace and Innocent Passage: Regulations for the Security …
23
Table 1 Actors, interests and objectives Space
Concerns
Actor
Objective
C Y B E R S P A C E
Sovereignty
State Organizations Hackers
Denying the dominance of information
Weapons Communications
Launching malicious software Blocking of telecommunications systems
Information
Spying through reconnaissance satellites
Infrastructure
Damage to critical infrastructure
Services
Attacking maritime traffic and control systems
Devices
Launching remote military attacks Maritime autonomous surface ship
Aircraft
Use cybernetic channel-controlled unmanned aircraft systems
Fishing
Altering identification systems (AIS)
Hydrography
Obtain information for the production of nautical charts Detect location of underwater minerals, marine ecosystems, shipwrecked species Identify updated routes
Research
Acquiring knowledge by infiltrating information systems and networks
Contamination
Cause environmental damage by attacking cargo and dangerous goods control systems
Advertising
States/ Organizations Activists Terrorists
Sponsor cyber-attacks Cause fear in networks Destroy data
Products
Offenders
Fraudulent freight forwarding
Others
Opportunists
Breaking cybersecurity defenses Make a profit
Source Prepared by the authors, based in [9]
employed by NATO, are used primarily to degrade or deceive an adversary’s command and control systems, nullifying their ability to make effective decisions while protecting one’s own and friendly command and control systems. Equally, merchant marine operations and navigational resources on board ships are not exempt from a cyber-attack during their innocent passage through the territorial sea. Another relevant aspect in this matter is the low earth orbit reconnaissance satellites, which provide high-resolution photographs fundamental in the planning of naval operations. The control and launching of military devices from anywhere in the world make it difficult to identify the origin of the attack and consequently hinders the right to legitimate defense of the State and its critical infrastructure. It also affects the security of international maritime traffic passing through the territorial sea. Consequently, this type of act can affect the national economy.
24
F. Ramírez-Cabrales and J. A. Machado Jiménez
Regarding the hydrography, scientific research, and fisheries, it is identified that databases and information resources on these matters can be infiltrated in order to steal information of national interest related to the geographical characteristics of the environment and the type of resources that lie in the territorial sea. In the same line, fisheries use automatic identification systems known as AIS. In this regard, criminals can access AIS global information systems to know the location of vessels. This is in order to launch viruses that hinder the existing coordination between the multiple naval units that operate command and control networks and logistics networks. Furthermore, it is noted that pollution of the marine environment can be induced and controlled by unauthorized access to loading and transport systems of dangerous substances on board ships. This can lead to unwanted natural disasters and irreversible damage to the territorial sea and its strategic ecosystems. On the other hand, it is interesting to broadcast propaganda for political purposes or to the detriment of the State’s national interests. Cyberspace can become a platform for media attention induced by State-sponsored organizations and activists interested in affecting the reputation of the State. Similarly, terrorists can generate fear in networks for political gain. As far as products are concerned, criminals can alter cargo records in a fraudulent manner. At the same time, they may be interested in commercial espionage to sell information and data. In the same vein, there are opportunists who can take advantage of cyberspace to breach cybersecurity systems in critical infrastructures for profit. Finally, there are no specific international instruments to regulate activities that may be considered harmful to the peace and security of the coastal State with regard to the activity of the right of innocent passage of foreign vessels through the territorial sea. In this regard, customary international law is limited exclusively to identifying activities that may affect the security of the State in that maritime space. At the same time, it empowers it to issue laws and regulations concerning innocent passage through the territorial sea. In general, there are mechanisms and documents that international society has encouraged to guide the legal actions of States. First, the International Maritime Organization, through the Committee on Safety of Navigation, has issued guidelines and circulars to ensure cyber security on board ships, and the deadline for compliance by the shipping company on January 1, 2021. Additionally, the literature records initiatives such as the International Telecommunications Union (ITU) Cybercrime Legislation Toolkit seeks to provide a framework for States to create effective legislation to address these crimes. Others such as the Commonwealth Model Law on Cybercrime, which like the Budapest Convention, contains provisions on criminal, procedural, and international cooperation material [24]. In addition, the Tallinn Manual on the International Law of Cyber Warfare is noteworthy. This document contains 95 rules, grouped into two parts: The Security of Cyberspace in International Law, and The International Law of Cyber Conflict. This is articulated throughout seven chapters. Among its objectives are contents on international law applied to cyberspace and the Net. With the aim of achieving consensus among States for an understanding of the application of resources in the framework of cyber defense and cybersecurity. It primarily highlights their intention
Cyberspace and Innocent Passage: Regulations for the Security …
25
Table 2 Criteria for determining cyber operations harmful to the security of the coastal State in its territorial sea Criteria
Determination
Classification
Severity
Cause physical damage
Harmful
Immediacy
Produce imminent damage
Harmful
Directness
Associate illegal behaviors
Harmful
Invasiveness
Interfere State interests
Harmful
Measurability of effects
Generate effects similar to the use of force
Harmful
Military nature
Articulate military and cyber operations
Harmful
State participation
Establish links between a State and cyber operations
Harmful
Alleged legality
Articulate military and cyber operations
Harmful
Source Prepared by the authors, based in [18]
to build an ethical consensus in this field and to mark the limits of what has been understood in this area as cyberwar, armed aggression, and the use of force [25]. In this context, the question that arises for international law is how it is determined whether a cyber-operation can constitute a threat to the peace and security of the coastal State during the innocent passage of foreign ships through the territorial sea. For this purpose, the analysis model proposed by Schmitt [26] was considered. This author proposes guidelines that allow States to identify tools to determine if a cyberoperation has reached the parameters to be considered a use of force. This criterion has been adopted by the Tallinn Manual in rule 11. This approach is useful because it allows to define when there is a use of illicit force in the field of cyber operations. The criteria to determine whether the foreseeable consequences of such an activity can affect the peace and security of the coastal State in its territorial sea respond to six factors indicated in Table 2. These factors are not absolute. This, taking into account the characteristics of this space, particularly its intangibility, makes its appropriation impossible and as a result makes it impossible for a State to exercise sovereignty in this area [27]. However, despite the fact that the State does not have sovereignty over cyberspace per se, it does have sovereignty over cyber infrastructure and over any activity that is related to these infrastructures, since these are over its territory. Therefore, these criteria constitute a first approximation to establish the guidelines that contribute to regulate the use of cyberspace to guarantee the security of the coastal State in its territorial sea and the activities associated with the right of innocent passage.
5 Conclusions The maintenance of the peace and security of the coastal State and the advancement of new information technologies require a reconsideration of the use of cyberspace and its impact on activities related to the right of innocent passage of foreign vessels
26
F. Ramírez-Cabrales and J. A. Machado Jiménez
over the territorial sea. This is in accordance with the fact that customary international law in no waypoints to the fifth element called cyberspace. Therefore, international law should review the issue of sovereignty and the jurisdiction applied to cyberspace. These elements are essential when determining the powers of the State to regulate cyber operations linked to the right to innocent passage through the territorial sea. Regarding sovereignty, it is demonstrated that, despite the impossibility of exercising unilateral domination of cyberspace, this does not preclude a State from exercising its authority in this area. Specifically, at least 15 national interests associated with the innocent passage of ships and their interactions with cyber operations that jeopardize the security of the coastal State were identified. Therefore, the imminent need to regulate these activities from the sovereign power of the State is inferred. For this purpose, international organizations, conventions, manuals, and initiatives that study cybercrime issues and their treatment from the judicial sphere were registered. These sources of law constitute inputs for the regulation of activities associated with the innocent passage and their interaction with cyber operations. In this regard, the Tallinn Manual was highlighted as a non-binding academic document that describes how to interpret international law in the context of cyber operations and cyber warfare. This manual provides six criteria, not exhaustive. Under these guidelines, the State can evaluate and consider the hostile behaviors of various actors who intend to generate cyber-attacks. Therefore, the document is a valuable contribution that can influence the approaches and positions of States and organizations on how to deal with this problem. As far as the regulation of the innocent passage is concerned, the document constitutes a reference of obligatory consultation to determine the legal regime of cyberspace—v. g. if it is a space that can be appropriated or if a particular legal regime is established to determine State competences both on the space itself considered and on the activities related to the innocent passage. In this way, it will be feasible to determine the responsibility of the State regarding the prosecution of criminal activities that take place in cyberspace. However, it should be noted that the regulation of innocent passage focuses exclusively on the activities of ships as a factor of instability for the security of the coastal State. This indicates that the international regime of maritime spaces in terms of regulating the right of innocent passage of ships does not respond to the interests and situations of conflict that may originate in cyberspace. This implies the need to seek an international consensus on the part of States in order to establish limits to the actions and interests of this new scenario. However, the rational condition of States in guaranteeing their survival in the international environment and maintaining the power of information delegitimizes any intention to create an international regime for the governance of cyberspace. Consequently, the trend is for each State in its sovereign capacity to be concerned with its own security and to issue laws and regulations that will make it possible to identify the types of activities that violate national sovereignty in controlling the security of international maritime traffic in the territorial sea and its undeniable relationship to the use and exploitation of cyberspace for the maintenance of national interests.
Cyberspace and Innocent Passage: Regulations for the Security …
27
References 1. Ocon, A.L., Gastaldi, S.: Ciberespacio y Defensa Nacional: Una Reflexión Sobre El Dilema Libertad-Seguridad En El Ejercicio De La Soberanía 1, p. 89. Revista Científica de la UNDEF (2019) 2. Dalbello, M.: Digital convergence: the past in the present. In: Spence Richards, P., et al. (eds.) A History of Modern Librarianship: Constructing the Heritage of Western Cultures. Libraries Unlimited Ed., California (2015) 3. Valeriano, B., Jensen, B., Maness, R.: Cyber Strategy. The Evolving Character of Power and Coercion. Oxford University Press, Oxford (2018) 4. Zekos, G.I.: State cyberspace jurisdiction and personal cyberspace jurisdiction. Int. J. Law Inform. Technol. 15, 1 (2007) 5. Ottis, R., Lorents, P.: Cyberspace: definition and implications. In: Proceedings of the 5th International Conference on Information Warfare and Security, pp. 267–270. Academic Publishing Limited (2010) 6. Joyanes, L.: Cibersociedad, Los retos sociales ante un nuevo mundo digital. Ed. McGraw-Hill (1997) 7. Nye, J.S.: Cyber Power, p. 94. Belfer Center for Science and International Affairs, Cambridge (2010) 8. U.S. Cyber Comand: Beyond the Build: Delivering Outcomes Through Cyberspace. Maryland. https://www.hsdl.org/?view&did=787006 (2015). último acceso 2019/12/11 9. Organización Marítima Internacional: Directrices sobre ciberseguridad a bordo de los buques, MSC 96/4/1. London. Homepage: http://www.imo.org (2016). último acceso 2019/12/11 10. Organización Marítima Internacional: Guidelines on Maritime Cyber Risk Managment. MSC-FAL 1/Circ.3. OMI, London. Homepage: http://www.imo.org (2017). último acceso 2019/12/11 11. Organización Marítima Internacional: Directrices del sector sobre ciberseguridad a bordo de los buques (versión 3). MSC 101/4/1. London. Homepage: http://www.imo.org (2019). último acceso 2019/12/11 12. Robles Carrillo, M.: El ciberespacio: Presupuestos para su ordenación jurídico-internacional, en Revista Chilena de Derecho y Ciencia Política, 7 que cita a RYAN et al. (2010): “International Cyberlaw: A Normative Approach”. Georget. J. Int. Law 42(2010–2011), 1161–1197 (2016) 13. Satola, D., Judy, H.: Towards a dynamic approach to enhancing international cooperation and collaboration in cybersecurity legal frameworks: reflections on the proceedings of the workshop on cybersecurity legal issues at the 2010 United Nations internet governance forum. William Mitchell Law Rev. 37(4), 1744–1804 (2010) 14. Hakapää, M., Molenaar, E.J.: Innocent passage: past and present. Mar. Policy 23(2), 131–145 (1999). https://doi.org/10.1016/S0308-597X(98)00032-3 15. UIT, Rec. UIT-T X.1205: Sector de Normalización de las Telecomunicaciones de la UIT (04/2008). Serie X: Redes de Datos, Comunicaciones de Sistemas Abiertos y Seguridad. Seguridad en el ciberespacio—Ciberseguridad. Aspectos generales de la ciberseguridad. Ginebra-Suiza (2009) 16. Bejarano, M.J.C.: Alcance y ámbito de la seguridad nacional en el ciberespacio. Cuadernos estrateg. 149, 47–82 (2011) 17. Raboin, B.: Corresponding evolution: international law and the emergence of cyber warfare. J. Natl. Assoc. Adm. Law Judic. 31, 2 (2011) 18. Llorens, M.P.: Los desafíos del uso de la fuerza en el ciberespacio. Anu. mex. der. int. 17, 785–816 (2017) 19. Kanuck, S.: Sovereign discourse on cyber conflict under international law. Texas Law Rev. 88, 1571–1597 (2010) 20. Shackelford, S.J.: From nuclear war to net war: analogizing cyber attacks in international law. Berkeley J. Int. Law 27(1), 193–251 (2009) 21. Vergara, M.D.A.: Ciberataques: La influencia del individuo como actor no estatal en riesgo a la seguridad nacional. Rev. Chil. Relac. Int. 3(55) (2019). ISSN 0719-8256
28
F. Ramírez-Cabrales and J. A. Machado Jiménez
22. Schmitt, M.N. (ed.): Tallinn Manual on the International Law Applicable to Cyber Warfare. Cambridge University Press (2013) 23. Ferrer, J.A.: Incidencia de la ciberguerra en las operaciones navales. In: Cuadernos de pensamiento naval: Suplemento de la revista general de marina, no. 15, pp. 87–102 (2013) 24. Loredo González, J.A., Ramírez Granados, A.: Delitos informáticos: su clasificación y una visión general de las medidas de acción para combatirlo. Celerinet, 44–51 (2013) 25. Domínguez, J.: La ciberguerra como realidad posible contemplada desde la prospectiva. Rev. Pensamiento Estratég. Seguridad CISDE 1(1), 18–32 (2016) 26. Schmitt, M.N.: Cyber operations and the jus ad bellum revisited. Villanova Law Rev. 56, 578–581 (2011) 27. Jensen, E.T.: Cyber sovereignty: the way ahead. Texas Int. Law J. 50(2), 275–304 (2015)
Security Versus Usability in E-Government: Insights from the Literature Fernando Huamán Monzón, Manuel Tupia, and Mariuxi Bruzza
Abstract Most governments establish regulations for the development of digital e-government services, aimed at improving the citizen’s experience and seeking to protect the information handled by those services. The problem is that these regulations do not contemplate a balance between information security and usability of web and mobile services; being that many times, both aspects are ignored in software design considerations. Achieving this balance serves as a component to measure the level of maturity of e-government services, so identifying that both elements are included in the development of web and mobile services is essential to guarantee the success of the joint implementation of e-government and compliance with the aforementioned regulations. In the present study, the different models for measuring the level of balance between safety and usability have been revisited as a review of the state-of-the-art. For this reason, the principles of PICO focused on the Population, Intervention, Comparison, and Outcomes have been used. The objective of the research is to identify those models that really take stock of both elements, or if they are only ad hoc efforts at the time of making specific developments of e-government applications. Keywords E-government · Information security · Usability · Maturity level model
F. H. Monzón · M. Tupia (B) Department of Engineering, Pontificia Universidad Católica del Perú, Av. Universitaria 1801, San Miguel, Lima, Peru e-mail: [email protected] F. H. Monzón e-mail: [email protected] M. Bruzza Universidad Laica “Eloy Alfaro” de Manabí, Ciudadela Universitaria S/N, Manta, Manabí, Ecuador e-mail: [email protected] © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_3
29
30
F. H. Monzón et al.
1 Introduction At present, many countries are implementing a series of electronic government services at different levels (country, department, municipality) [1]. One of the most challenging hurdles faced by the electronic government implementation projects is supplying a proper level of information security [2] to such a degree that now this is considered a crucial factor to reach an advanced level in electronic government [3]. Usability is seen as a catalytic and success factor for the electronic government; it is even related to the population and its per capita income [4]. Periodic evaluations of usability are conducted for the implementation projects of electronic government, whereas the inclusion of other technical aspects (such as the infrastructure, terminals, suppliers, etc.) and social ones (social state, demographic parameters, etc.) is under evaluation [5]. That is the reason why several authors claim that usability and security must be both considered when designing new information technology [6–8]. This paper is organized as follows: Sect. 2 introduces a brief description of concepts, the pillar of this research: security, usability, and electronic government; Sect. 3 deals with PICOC based systematic review process and shows the state-of-the-art results in a summarized and organized manner; Sect. 4 introduces the discussions on these results; Sect. 5 provides conclusions and future work based on the review.
2 Background on Security and Usability in the Electronic Government The electronic government is defined as the adoption of ICTs in the public administration structures such as channels interconnecting and interacting with organizations and people through webs, electronic mail, mobile applications, intranet, extranet, etc. [9]. A similar concept is established by the Organization of American States (OAS), including the objective of the technology used by the government, i.e., qualitatively improving services and information available to citizens, increasing the efficiency and efficacy of the public managements and improving substantively transparency of the public sector and citizens participation [10]. Furthermore, information security is defined as the set of processes and activities allowing maintaining information assets (part of an organization) free of dangers and hazards due to incidents or attacks [11]. Today information (physical, printed, digital) has become a key resource for organizations [12]. The concept of security has also been applied to information, with information security being defined as the preservation of confidentiality, integrity, and availability of information [13]. Usability is the characteristic of systems that is known as “the degree in which a product may be used by specific users to achieve specific objectives in a context of specific use” [14]. It is also established as an attribute in terms of software quality, which assesses how easy to use user interfaces may be. Usability is defined in terms
Security Versus Usability in E-Government: Insights from …
31
of five attributes: learning, efficiency memory, prevention of errors, and subjective satisfaction [15]. Today, usability and information security are recognized as a requirement in the electronic government services [16, 17]. They are also considered success factors for the services design [18], as well as the implementation of the electronic government services [19–22].
3 State-of-the-Art Systematic Review Databases to be used for the systematic review are the following: SCOPUS, ACM, IEEE, and Springer. Table 1 shows PICOC criteria [23] that can be used in the systematic review. The comparison criterion is not applicable because the systematic review purpose is to evidence the main progress in researches, and not compare such results.
3.1 Research Questions The objective of this research is to conduct a mapping of literature in which researchers have related information security to usability in electronic government services. This leads to the following research questions. Table 2 details research questions for a systematic review. Table 1 Attributes according to PICOC criteria Heading level
Example
Population
Information security and usability models
Intervention
Designed, analyzed and/or applied models from 2010 to 2020
Context
In electronic government services
Outcomes
Information security versus usability models
Comparison
(Not applied)
Table 2 Research questions No.
Research questions
1
Are information security and usability part of the requirements in the implementation of electronic government services?
2
How information security relates to usability in electronic government services?
3
How to configure information security and usability in applications or specific solutions of electronic government?
32
F. H. Monzón et al.
Table 3 Search string Database
Search string
Filters
Total
Scopus
ALL [(“information security” OR security OR secure) AND (usability OR usable) AND (e-government OR e-gov)] AND PUBYEAR > 2009
2010–2019
1191
ACM
(“information security” OR security OR secure) AND (usability OR usable) AND (e-government OR e-gov)
Published since: 2010
98
IEEE
(“information security” OR security OR secure) AND (usability OR usable) AND (e-government OR e-gov)
Filters applied: 2010–2019
25
Springer
“(‘information security’ OR security OR secure) AND (usability OR usable) AND (e-government OR e-gov)”
2010–2019
Total
552
1866
3.2 Strategy and Search Strings For the search process of primary papers the following steps are shown below: 1. PICOC criteria were used as search terms and both synonyms and abbreviations have been included. 2. For the search chain Boolean operators have been used (OR, AND). The search chain used for the systematic review of literature has been customized to match the syntaxes of each database chosen as shown in Table 3.
3.3 Selection Criteria Inclusion and exclusion criteria were specified according to research questions as detailed in Table 2. For the final selection of papers in the review, the following inclusion criteria were taken into account: • Papers framed in electronic government. • Information security and usability are related to some extent. • Papers analyzing the father–son interaction between 1 of the 2 factors (information security, usability or usability in information security). • Papers evidencing the need for usability and information security for the electronic government services.
Security Versus Usability in E-Government: Insights from …
33
• Papers including security and usability as success factors in the implementation of technology services. • Security and usability models. • Study cases, applications or specific solutions of security and usability. • Security and usability assessments in technological services. On their side, the exclusion criteria would be the following: • Papers developing outside the electronic government context. • Papers to which access cannot be gained. • Papers mentioning terms such as security and/or usability (in abstracts, in keywords), but that are not eventually referred to.
3.4 Selection Process The study identified a total of 1866 primary papers. After verifications, 1812 articles were ruled out because they did not comply with the inclusion criteria, whose summary is shown in Table 4. The remaining 54 papers are the studies approved for this review, with an extract thereof seen in Table 5. Table 4 Selected papers Database
Total papers
Selected papers
Scopus
1191
38
98
3
25
13
ACM IEEE Springer Total
552
0
1866
54
Table 5 Selected process Paper title
Reference
Publication year
Identifying indexes affecting the quality of e-government websites
[24]
2019
Information security risk assessment model based on OCTAVE for e-government
[25]
2010
Information security strategy on mobile device based e-government
[26]
2015
LBACWeb: a lattice-based access control model for mobile thin client based on web OSes
[27]
2019
LeMTrac: legislative management and tracking system
[28]
2019
34
F. H. Monzón et al.
3.5 Results of the Systematic Review Based on the research questions that were previously posed and following the PICOC methodology, the results obtained from the selected papers are introduced below, classified by a question. Furthermore, for the sake of order and grouping the following results organization has been generated and introduced below, and they are covered in the Discussion topic (Table 6). Results of Question 1 For question 1 Are information security and usability part of the requirements in the implementation of electronic government services? a total of 23 papers were identified and then classified into 3 groups. P1G1—Need in the Electronic Government. The first group P1G1 evidences information security and usability as critical requirements for electronic government services. Table 7 lists the results. P1G2—Electronic Government Evaluation. Table 6 Research questions Research question
Group
P1—Are information security and usability part of the requirements in the implementation of electronic government services?
P1G1—Need in the electronic government P1G2—Electronic government evaluation P1G3—Electronic government success factors
P2—How information security relates to usability in electronic government services?
P2G1—Security and usability relation
P3—How to configure information security and usability in applications or specific solutions of electronic government?
P3G1—Technological applications
Table 7 Answers to question 1a Research question
Group
E-government transparency and citizen engagement increasing accountability
A study case of electronic government web services so that they are usable in a secure manner
An adoption model of electronic government services in Malaysia: electronic labor exchange (ELX)
Identifies an urgency of electronic government services improvement in terms of usability and security
Multipath routing slice experiments in federated testbeds
Analysis of Internet towards the future considers security, quality and reliability as requirements in the coming years. Usability takes this as know-how of the work team
Proceedings of the 5th International Symposium on Human Aspects of Information Security and Assurance, HAISA 2011
Evidence suggests that there is a need for a novel approach to improve perception and usability of information security
Security Versus Usability in E-Government: Insights from …
35
The second group P1G2 contains papers in which Electronic Government services quality evaluation has been conducted and security and usability have been taken into account as part of the parameters to be evaluated. It is relevant to point out that this group of papers evidences the significance of these two factors in electronic government services. Table 8 shows partial results. Table 8 Answers to question 1b Research question
Group
Exploring the dimensions of electronic government service quality
Proposes a scale to measure the dimension of electronic government services quality: (by literature review since the last 3 decades, and through the review of the SERVQUAL scale). Dimensions of quality: reliability, answer capacity, website design\contents, and security\privacy
An assessment study of Indian state government portals
Electronic government services quality based on the utility of information, adequacy of information, citizen-centric information, usability, accessibility, interaction, privacy, security, and citizen participation
Citizen participation through municipal websites: a global scorecard
The study assessed the municipal websites in five different categories of electronic governments: (1) security and privacy, (2) content, (3) usability, (4) services, and participation (5) of citizens. Seoul is the top-ranked city
E-service quality model for Indian government portals: citizens’ perspective
Quality perceived by users. Seven constructs, i.e., citizens centricity, transaction transparency, technical adequacy, usability, complete information, privacy and security, and usefulness of information—were identified from the analyses, which may be used to assess the service quality at the request of government portals
Assessment of development level of municipal websites of the Republic of Lithuania
The e-gov performance has been assessed taking into account usability and security, but no relation is established
Explaining the underdevelopment of rural e-government: the case of Romania
Assessment in Romania under five aspects: (1) security and protection of personal data; (2) usability; (3) content; (4) type of services; and (5) digital democracy
State e-government portals in Malaysia: an empirical investigation
The authors analyzed the website for a total of thirteen states in Malaysia in relation to six aspects: extent of transparency, interactivity, usability and portal accessibility, citizen participation, security and privacy, and the maturity level of service. They present a roadmap to improve these aspects
36
F. H. Monzón et al.
Table 9 Answers to question 1c Research question
Group
e-Participation initiatives in Europe: learning from practitioners
Determine success factors in e-Participation: government’s commitment; usability; online combination with off-line channels; a complete communication and promotion plan; security and privacy; organizational aspects; and complexity and quality of e-Participation
Evaluating Jordan’s e-government website: a case study
Identifies that usability, accessibility, and privacy/security are key factors for the successful adoption of electronic government webpages
Secure and reliable online verification of electronic signatures in the digital age
Among other things, security, usability and privacy preservation may be identified as key components for the verification of reliable electronic signature
Citizens’ attitudes towards electronic identification in a public e-service context—an essential perspective in the eID development process
By means of focus groups, usability and security are identified as key factors for the success of electronic government services, in a secure authentication. Citizen’s attitudes have shown they constitute an important addition to the development of electronic identification solutions supporting the success of electronic government
Designing e-government services: key service attributes and citizens’ preference structures
Four key factors were identified for the adoption of transactional services in e-gov: usability, computer resources, technical assistance, and security
P1G3—Electronic Government success factors. The third group P1G3 considers papers that have included security and usability as part of the success factor for the Electronic Government services. Table 9 shows partial results. Results of Question 2 Question 2 How information security relates to usability in electronic government services? intends to identify those researches proposing a way to link both factors at the same level or in the father–son relationship. Table 10 lists some of the papers providing an answer to this question 2. Results of Question 3 Question 3 How to configure information security and usability in applications or specific solutions of electronic government? intends to identify those applications or technological solutions which involved security and usability requirements (by each context or stage). This group aims at finding out manners in which the relation of both factors can materialize within an electronic government solution. Table 11 shows another extract of papers answering this question.
Security Versus Usability in E-Government: Insights from …
37
Table 10 Answers to question 2 Research question
Group
Information security risk assessment model based on OCTAVE for e-government
Information security risk model for electronic government considers the confidentiality, integrity of information, and usability to define the value of information assets
Mobile government for improved public service provision in South Africa
A model to develop electronic government services considering security and usability (other factors: great impact, sustainable, and development)
Factors supporting user interface design of mobile government application
A total of eleven factors that show correct usability are identified, security is among them
LeMTrac: legislative management and tracking system
A study evidencing the good acceptance of electronic government services in terms of usability, and the enhancement of this acceptance when accessibility, security and document management are added
4 Discussion As a result of the Systematic Review, 54 papers were identified, and then organized and classified as follows: In group P1G1 (4 papers)—Need in the Electronic Government—electronic government services showed that they need to be usable and secure at present [29], because the future trend asks for these two factors because, for example, citizens’ personal information is involved [16, 17, 30]. In group P1G2 (13 papers)—Electronic Government Evaluation—security and usability were discussed among the evaluation variables. Government websites [50], electronic government services at municipal level [31, 32] and assessments under the SERVQUAL model [33] were assessed. In addition, electronic government assessments were conducted in India [18, 34], Lithuania [35], Romania [36, 37], Malaysia [38], Kyrgyzstan [39], Oman [40], and a global assessment [41] as well. In group P1G3 (9 papers)—Electronic Government success factors—security and usability were identified as success factors for the services design [18], optimization [42], user-focused designs [43], as well as the implementation of electronic participation [19], electronic voting [20], electronic signature [21], and electronic identification [22], and in specific scenarios such as in Saudi Arabia [44] and Jordan [45]. In group P2G1 (4 papers)—Security and usability relation—risk models [25] and models for the development of mobile electronic government services [46] were identified including certain security and usability aspects as part of their components. In addition, two papers including security as a factor to improve usability [28, 47] as a father–son relationship since it seeks with an independent factor to improve the dependent one.
38
F. H. Monzón et al.
Table 11 Answers to question 3 Research question
Group
LBACWeb: a lattice-based access control model for mobile thin client based on web OSes
Proposes a special configuration in LBACWeb to increase its usability
Towards user-friendly e-government solutions: usability evaluation of Austrian smart-card integration techniques
Security and usability combination for authentication: MOCCA-local and MOCCA-online. Both are secure but each fulfills different usability requirements
Measuring usability to improve the efficiency of electronic signature-based e-government solutions
A study directly relating security and usability in authentication. It shows the results of the study
Towards cross-domain eID by using agile mobile authentication
About a more usable secure authentication and identification method
Biometrics privacy: technologies and applications
An analysis on biometric authentication, security of stored biometric information, and usability in this type of authentication
Mobile-only solution for server-based qualified electronic signatures
Improves biometric authentication with QES, increasing security, and usability
Research on governmental data sharing based on local differential privacy approach
A private blockchain enhances security, credibility, and capacity of information exchange response among the state departments. Local differential privacy provides more usability and security for the exchange of statistics. Not only maintains available statistics, but also protects citizens’ privacy
E-government services: Italian certified electronic mail
A certified email in Italy, as a success case of combining security and usability
User perception of Bitcoin usability and security across novice users
Within cryptocurrency use acceptance background, they are regarded as secure but need to boost their usability with respect to users
Optimization of digitalized document verification using e-governance service delivery platform (E-SDP)
UDI (Unique Document Identifier) issued by the government to improve the verification mechanism of security, verifiability, usability, to control the state, Aadhaar allowed access, and authenticity of the document
Breaking the barriers of e-participation: the experience of Russian digital office development
A study case of digital signature in Russia: authentication system (usability and security)
Usability evaluation of electronic signature-based e-government solutions
A study relating security and usability in the Austrian government. Components assessed to perform as middleware and facilitate integration of cryptographic hardware tokens in the electronic government applications
Usability evaluation of electronic signature-based e-government solutions
A study relating security and usability in the Austrian government. Components assessed to perform as middleware and facilitate the integration of cryptographic hardware tokens in the electronic government applications
Security Versus Usability in E-Government: Insights from …
39
For group P3G1 (24 papers)—Technological applications—solutions relating security and usability in authentication aspects [48–50], including biometric authentication [51, 52], access management [27], as well as identity management [53–55], mobile identity [56, 57] and electronic identity [58, 59] digital signature [60] and electronic signature [61–63] solutions were found. Applications in mobile services [26] and specific technology such as the use of a unique identification document [64], Single Sign-On [21, 65], certified electronic mail [66], blockchain [67] and cryptocurrency [68] were identified. This shows that researchers and engineers have required to combine both factors for the design of such technological solutions, thus demonstrating the significance of security in the information to be managed by technology and usability with respect to the end user.
5 Conclusions and Future Works The systematic review of the state-of-the-art reaffirms the importance of incorporating information security and usability within the requirements of e-government digital services development. Many works indicate that they are going to refer to the balance of safety and usability, but finally, they prefer one of the two factors. This is one of the reasons why the amount of useful papers for the present investigation is drastically reduced. It is concluded, from the studies reviewed, that this balance is not a general practice in e-government implementations but is included in a reactive manner, in the face of problems of any kind. Likewise, it is concluded that one of the factors for the evaluation of the quality of e-government services is the presence of this balance. Some frameworks or models of maturity include it as the main factor such as [25, 46]. It is also concluded that in Asia (Saudi Arabia [44], Jordan [45], India [18, 34], Malaysia [38], Kyrgyzstan [39] and Oman [40]), and in Europe (Lithuania [35], Romania [36, 37]) have developed the most successful web and mobile applications where the interaction between security and usability is natural. Within this context, it is important to mention that there are very few investigations carried out in South America, and more specifically Peru and Ecuador. As future work, it is proposed to start the construction of a model that balances security and usability in government electronic services, both in the design and implementation steps.
40
F. H. Monzón et al.
References 1. Nixon, P.G., Koutrakou, V., Rawal, R.: Understanding E-Government in Europe. Routledge, London (2010) 2. Alguliev, R., Imamverdiyev, S.: e-Government gestión de seguridad de la información desafíos de investigación. Bakú, Azerbaiyán, s.n. (2009) 3. Wang, J.-F.: E-Government Security Management: Key Factors and Countermeasure (2009) 4. Youngblood, N.E., Youngblood, S.A.: User experience and accessibility: an analysis of county web portals. J. Usabil. Stud. 9(1), 25–41 (2013) 5. Nariman, D.: Assessing the Impact of E-Government on Users: A Case Study of Indonesia. Palermo, Italy (2012) 6. Cranor, L.F., Buchler, N.: Better together: usability and security go hand in hand. IEEE Secur. Priv. 12(6), 89–93 (2014) 7. Lassmann, G.: Some Results on Robustness, Security and Usability of Biometric Systems. Lausanne, Switzerland, Switzerland, s.n. (2002) 8. Ferreira, A., Rusu, C., Roncagliolo, S.: Usability and Security Patterns. Cancun, Mexico (2009) 9. Criado, J.I., Ramilo, M.D.C.: ¿Un reto o una nueva moda? Problemas y perspectivas de futuro en torno a Internet y las tecnologías de la información y la comunicación en las administraciones públicas del siglo XXI (2001) 10. OEA: Organización de los Estados Americanos: Democracia para la paz, la seguridad y el desarrollo. http://www.oas.org/es/cidh/indigenas/proteccion/cautelares.asp (2009) 11. Bosworth, S., Kabay, M.E.: Computer Security Handbook. Wiley. https://dl.acm.org/citation. cfm?id=560494 (2002) 12. ISACA: COBIT® 5 for Information Security (2012) 13. ISO: Information Technology—Security Techniques—Information Security Management Systems—Overview and Vocabulary. ISO/IEC 27000:2018 (2018) 14. ISO: ISO 9241-11:2018—Ergonomics of Human-System Interaction—Part 11: Usability: Definitions and Concepts (2018) 15. Nielsen, J.: Usability Engineering (1993) 16. Noor, Z.M., Kasimin, H., Aman, A., Saharai, N.: An adoption model of electronic government services in Malaysia: electronic labor exchange (ELX). J. Pengurusan, 87–97 (2011) 17. Zaaba, Z.F., Furnell, S., Dowland, P.: Usability and Security Patterns. London, UK, s.n. (2011) 18. Bhattacharya, D., Gulla, U., Gupta, M.: An assessment study of Indian state government portals. In: E-Government Website Development: Future Trends and Strategic Models, pp. 130–152. IGI Global, India (2010) 19. Panopoulou, E., Tambouris, E., Tarabanis, K.: eParticipation Initiatives in Europe: Learning from Practitioners. Lausanne, Switzerland (2010) 20. Alkhakani, S.B., Hassan, S.M.: Suggest trust as mediation to adopt electronic voting system in Iraq. J. Theor. Appl. Inf. Technol. 96(17), 5729–5739 (2018) 21. Zefferer, T., Zwattendorfer, B., Tauber, A., Knall, T.: Secure and Reliable Online-Verification of Electronic Signatures in the Digital Age. Rio de Janeiro, Brazil (2011) 22. Axelsson, K., Melin, U.: Citizens’ Attitudes Towards Electronic Identification in a Public EService Context—An Essential Perspective in the eID Development Process. Kristiansand, Norway (2012) 23. Brereton, P., Kitchenham, B., Budgen, D., Turner, M., Khalil, M.: Lessons from applying the systematic literature review process within the software engineering domain. J. Syst. Softw. 80, 571–583 (2007) 24. Shayganmehr, M., Montazer, G.A.: Identifying Indexes Affecting the Quality of E-Government Websites. Tehran, Iran, Iran, s.n. (2019) 25. Chen, X., Wen, N.: Information Security Risk Assessment Model Based on OCTAVE for E-Government, pp. 1–5. IEEE, Wuhan, China (2010) 26. Priyambodo, T.K., Prayudi, Y.: Information security strategy on mobile device based e-government. ARPN J. Eng. Appl. Sci. 10(2), 652–660 (2015)
Security Versus Usability in E-Government: Insights from …
41
27. Yang, Y., Song, Z.: LBACWeb: A Lattice-Based Access Control Model for Mobile Thin Client Based on Web OSes, pp. 103–109. Kuala Lumpur, Malaysia (2019) 28. Pandes, T.L.O., Omorog, C.D., Medrano, R.B.: LeMTrac: Legislative Management and Tracking System. IEEE, Baguio City, Philippines (2018) 29. Carstens, D.S., Kies, S., Stockman, R.: E-government transparency and citizen engagement increasing accountability. In: Information Resources Management Association (USA) (ed.) Digital Solutions for Contemporary Democracy and Government, pp. 189–214 (2015) 30. Zseby, T., Zinner, T., Tutschku, K., Shavitt, Y., TranGia, P., Schwartz, C., Rafetseder, A.: Multipath routing slice experiments in federated testbeds. In: The Future Internet Assembly, FIA 6656, pp. 247–258 (2011) 31. Schatteman, A., Mohammed-Spigner, D., Poluse, G.: Citizen participation through municipal websites: a global scorecard. In: Active Citizen Participation in E-Government: A Global Perspective, pp. 403–414. Information Science Reference, Illinois, USA (2012) 32. Manoharan, A., Zheng, Y., Melitski, J.: Global comparative municipal e-governance: factors and trends. Int. Rev. Publ. Adm. 22(1), 14–31 (2017) 33. Balushi, T.H.A., Ali, S.: Exploring the Dimensions of Electronic Government Service Quality, s.l., s.n. (2016) 34. Bhattacharya, D., Gulla, U., Gupta, M.: E-service quality model for Indian government portals: citizens’ perspective. J. Enterp. Inf. Manag. 25(3), 246–271 (2012) 35. Domarkas, V., Laukaityte, A., Maˇciukas, V.: Assessment of development level of municipal websites of the Republic of Lithuania. Public Policy Adm. 11(1), 23–36 (2012) 36. Stoica, V., Ilas, A.: Explaining the underdevelopment of rural e-government: the case of Romania. In: E-Government Implementation and Practice in Developing Countries, pp. 331–348 (2013) 37. Stoica, V., Ilas, A.: An assessment of rural e-government in Romania. In: E-Government Implementation and Practice in Developing Countries. Como, Italy, s.l.:s.n. (2013) 38. Eskandar, A.A., Raman, M.: State e-government portals in Malaysia: an empirical investigation. Int. J. Electron. Gov. Res. 9(2), 19–46 (2013) 39. Ismailova, R.: Web site accessibility, usability and security: a survey of government web sites in Kyrgyz Republic. Univ. Access Inf. Soc. 16(1), 257–264 (2017) 40. Bulushi, T.A.: Quality of E-Government Services in Oman: A Perspective of E-Government Services Providers. Spain, Madrid (2017) 41. Holzer, M., Manoharan, A., Fudge, M.: Global trends in digital governance: a longitudinal study. Int. J. Technol. Diffus. 2(2), 32–46 (2011) 42. Li, Z.H., Wu, Y.X., Xu, S.H., Zhang, W.: E-Government Application Adaptation and Optimization Based on the Domestic Foundational Software. Zhuhai, China, s.n. (2013) 43. K˝o, A., Molnár, T., Mátyus, B.: A User-centred Design Approach for Mobile-Government Systems for the Elderly. Phnom Penh, Cambodia, Cambodia, s.n. (2018) 44. Aljarallah, S., Lock, R.: An Empirical Study of Sustainable E-Government Characteristics in Saudi Arabia. Santiago de Compostela, Spain (2018) 45. Abu-Shanab, E., Baker, A.N.A.: Evaluating Jordan’s e-government website: a case study. Electron. Gov. 8(4), 271–289 (2011) 46. Nkosi, M., Mekuria, F.: Mobile Government for Improved Public Service Provision in South Africa. IEEE, Durban, South Africa (2011) 47. Kureerung, P., Ramingwong, L.: Factors Supporting User Interface Design of Mobile Government Application, pp. 115–119. Tokyo, Japan (2019) 48. Zefferer, T., Krnjic, V.: Towards User-Friendly E-Government Solutions: Usability Evaluation of Austrian Smartcard Integration Techniques. Vienna, Austria (2012) 49. Zefferer, T., Krnjic, V., Stranacher, K., Zwattendorfer, B.: Measuring usability to improve the efficiency of electronic signature-based e-government solutions. In: Public Administration and Information Technology, pp. 45–74. Springer, New York (2014) 50. Lenz, T., Alber, L.: Towards Cross-Domain eID by Using Agile Mobile Authentication. Sydney, NSW, Australia, s.n. (2017)
42
F. H. Monzón et al.
51. Piuri, V., Scotti, F.: Biometrics Privacy: Technologies and Applications. Seville, Spain, s.n. (2011) 52. Theuermann, K., Tauber, A., Lenz, T.: Mobile-Only Solution for Server-Based Qualified Electronic Signatures. Shanghai, China, China, s.n. (2019) 53. Neubauer, T., Heurix, J.: A Roadmap for Personal Identity Management. IEEE, Menuires, France (2010) 54. Alotaibi, S.J., Wald, M.: Towards a UTAUT-Based Model for Studying the Integrating Physical and Virtual Identity Access Management Systems in E-Government Domain. London, UK (2012) 55. Alotaibi, S.J., Wald, M.: Evaluation of the UTAUT Model for Acceptable User Experiences in Identity Access Management Systems. London, UK, s.n. (2013) 56. Verzeletti, G.M., de Mello, E.R., Wangham, M.S.: A National Mobile Identity Management Strategy for Electronic Government Services. New York, NY, USA, s.n. (2018) 57. Verzeletti, G.M., de Mello, E.R., Wangham, M.: A mobile identity management system to enhance the Brazilian electronic government. IEEE Lat. Am. Trans. 16(11), 2790–2797 (2018) 58. Zefferer, T., Teufl, P.: Leveraging the Adoption of Mobile eID and e-Signature Solutions in Europe. Spain (2015) 59. Zwattendorfer, B., Tauber, A.: Secure single sign-on authentication using eIDs across public clouds. Int. J. Internet Technol. Secur. Trans. 5(4), 291–306 (2014) 60. Gorelik, S., Lyaper, V., Bershadskaya, L., Buccafurri, F.: Breaking the Barriers of eParticipation: The Experience of Russian Digital Office Development. Munich, Germany (2014) 61. Zefferer, T., Krnjic, V.: Usability Evaluation of Electronic Signature Based E-Government Solutions. Madrid, Spain (2012) 62. Krnjic, V., Stranacher, K., Kellner, T., Fitzek, A.: Modular Architecture for Adaptable Signature-Creation Tools Requirements, Architecture, Implementation and Usability. Koblenz, Germany (2013) 63. Zefferer, T.: A Server-Based Signature Solution for Mobile Devices. New York, NY, USA, s.n. (2014) 64. Raghunathan, V., Raj, V.C., Eswaran, S., Ambika, A., Sangeetha, R.U.: Optimization of digitalized document verification using E-Governance service delivery platform (E-SDP). Int. J. Appl. Eng. Res. 11(4), 2535–2543 (2016) 65. Zwattendorfer, B., Tauber, A.: Secure Cross-Cloud Single Sign-On (SSO) Using eIDs. London, UK, s.n. (2012) 66. Buzzi, M., Gennai, F., Petrucci, C., Vinciarelli, A.: E-Government Services: Italian Certified Electronic Mail. Porto, Portugal (2014) 67. Liu, L., Piao, C., Jiang, X., Zheng, L.: Research on Governmental Data Sharing Based on Local Differential Privacy Approach. Xi’an, China, s.n. (2018) 68. Alshamsi, A., Andras, P.: User perception of Bitcoin usability and security across novice users. Int. J. Hum. Comput. Stud. 126, 94–110 (2019)
Simulation and Computer Vision in Military Applications
Ackermann UGV with 2D Mapping for Unknown Environments Wilbert G. Aguilar, Israel Asimbaya, Pablo Albán, and Yéssica Fernández
Abstract UGVs have been used to replace humans in high-risk tasks such as explore caverns, minefields or places contaminated with radiation but also, they are used to research path planners in order to replace human intervention to drive a car. UGVs can be controlled remotely from a safety place to avoid injuries, lethal damage and reduce fatal accidents. However, in order to acquire information about vehicle surroundings, it is necessary to use sensors or cameras that provide information to the remote operator in order to make the best decision of the vehicle course. Correspondingly, the information acquired by the sensors can be used to build an environment map, which will be useful for future applications, or to save a register of the explored area. The work proposed develop and build a vehicle with an Ackermann steering that can be controlled remotely by an operator using a portable computer, with the purpose of explore unknown environments using stereo vision cameras and build a map with information about the surroundings. Keywords UGV · Stereo vision · Remote control · Mapping
1 Introduction Nowadays, UGVs are used in investigations due to its utility at the time of replacing the human factor to execute high-risk tasks, such as exploring unknown environments, reconnaissance of dangerous routes or conflict area surveillance [1, 2]. W. G. Aguilar (B) · I. Asimbaya · P. Albán CICTE, DEEL, Universidad de las Fuerzas Armadas ESPE, 171103 Sangolquí, Ecuador e-mail: [email protected] W. G. Aguilar FIS, Escuela Politécnica Nacional, Quito, Ecuador GREC, Universitat Politècnica de Catalunya, Barcelona, Spain Y. Fernández Universidad Nacional de Educación, Azogues, Ecuador © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_4
45
46
W. G. Aguilar et al.
Researchers also improve UGV analysis, designing a 3D environment model where the vehicle can simulate navigation in a dangerous situation, like buildings near to collapse, also interaction with the surface can be estimated to improve simulation results [3, 4]. However, researchers seek to improve the performing of remote-control stations and its facility to manipulate the vehicle by adding better network devices for communication or simplifying graphic interfaces [5, 6]. According to [7, 8], there exist different types of modes to control UGVs, the first way is commanded control that depends on decisions of the operator based on information providing by sensor or cameras mounted on the vehicle; the second one is self-controlled that use a path planner to avoid obstacles and decide the best trajectory based on information provided by a GPS or by the sensors, gesture control, where navigation is commanded by values acquired by an IMU mounted in a glove which change its values in function of hand movement, and finally, raptor control, which make and advanced control tracking by image processing algorithms. Environment perception is used to identify obstacles and show 3D structures [9, 10], however, different technologies are used to identify the quality of ground [11, 12] or possible obstacles such as thermal lectures, where obstacles show a higher temperature than background with the advantage that can be used at night or environments with poor visibility [13, 14]. In order to increase data reliability, the number of sensors increases to support weakness of each other, as shown in [15, 16] where are mounted linear lasers, radar an infra-red camera and a color camera, which are synchronized to collect data while the vehicle is static or is moving. On the contrary, the amount of sensors also requires computational processing, for that reason the use of binocular and multi-camera vision is increasing in use, based on computer vision techniques to acquire environment data using disparity reconstruction [17, 18] that deliver object depth information in grayscale. This article proposes design and construction of an Ackermann steering vehicle that can be controlled remotely with wireless communication in order to generate a 2D map from cloud point data, using computer vision techniques to convert 3D information into 2D saturation map and use a stereo vision camera as the input sensor.
2 Our Approach 2.1 Cinematic Model of Ackermann Steering The Ackermann steering is composed of 4 bars mechanism connected to the front wheels as shown in Fig. 1, in order to change the vehicle’s trajectory [19]. The vehicle makes circular trajectories where the radius origin is out of the vehicle. The center can be found by intersecting the horizontal axis of back wheels and the perpendicular of the front wheels axis. As shown in Fig. 2, the center of the circular trajectory is a function of the angles from the two front wheels and the vehicle dimensions such as the length and width. To relate all this factors, Eq. (1) is defined.
Ackermann UGV with 2D Mapping for Unknown Environments
47
Fig. 1 Ackermann steering 4 bars mechanism. a Straight trajectory; b curved trajectory
Fig. 2 Relation between angles and vehicle dimensions
cot δ2 − cot δ1 =
W L
(1)
On the other hand, this model represents trajectory having as a reference one of the two front wheels, for that reason an equivalent wheel is defined as shown in Fig. 3, also it helps to simplify the mathematical model. The equivalent angle can be defined as shown in Eq. (2) and the Ackermann angle is shown in Eq. (3). cot δ =
cot δ2 + cot δ1 2
(2)
L R
(3)
δACK =
Now we can assume that the Ackermann angle is approach to the sum of the turning angles as shown in Eq. (4). δ1 + δ2 ≈ δ ≈ δACK
(4)
48
W. G. Aguilar et al.
Fig. 3 Equivalent middle wheel
2.2 Environment Perception The stereo camera uses stereoscopic vision that captures two images of the same scene, but is separated by a known distance, then the saturation points are identified in both images and finally, the points are matched [20]. The number of displacement pixels is directly related to the depth distance as shown in Fig. 4. The stereo camera delivers a cloud of points that describe the depth of the environment, each point has information of its position on the cartesian axis; however, Fig. 4 Stereoscopic vision explanation
Ackermann UGV with 2D Mapping for Unknown Environments
49
Fig. 5 2D map construction
in order to build a 2D map is necessary to normalize each point to a common “z” point, and the rest of the points are discarded. The algorithm read camera information constantly and also transforms the 3D information into 2D information to build the map as shown in Fig. 5. Finally, the map can be saved as a pgm image that has information on the environment based on saturation, where black points mean full saturation (obstacle) and white points mean low saturation or empty space, the result is shown in Fig. 6.
2.3 Remote Control The communication between the vehicle and the operator is established by a LAN network as shown in Fig. 7 where the laptop sends commands using ROS topics and nodes [20] to the processing card inside the vehicle, allowing the operator to control vehicle direction, the velocity of displacement and angle of curvature [21]. The code that receives the commands from the operator is shown below (Algorithm 1). To control the vehicle the program reads constantly input data each 100 ms in order to establish a real-time communication, where the operator sets the direction parameters, however, to visualize the vehicle perspective the processing card streams camera images that are visualized in the laptop used by the operator.
50
Fig. 6 Final map built
Fig. 7 UGV teleoperated vehicle communication diagram
W. G. Aguilar et al.
Ackermann UGV with 2D Mapping for Unknown Environments
51
3 Conclusions and Future Work This paper shows a UGV with Ackermann steering that can be controlled remotely by a LAN network using a laptop with a simple interface (HMI) that sends control messages using ROS nodes and topics to control actuators inside the vehicle in order to acquire information of the environment, process that information and building a 2D map, where the amount of sensors has been reduced to the minimum required to perceive the surroundings. Results show that the vehicle control is acceptable due to commands time response that also improves operator experience while the mapping is running, also the real-time transmission of camera information allows the operator to guide the vehicle through the best path. In addition, the built map shows the empty spaces and obstacles in the vehicle’s trajectory that can be used to develop a path planner for Ackermann vehicles, considering its kinematics restrictions. Acknowledgements This work is part of the project perception and localization system for autonomous navigation of rotor micro aerial vehicles in GPS-denied environments, VisualNavDrone, 2016-PIC-024, from the Universidad de las Fuerzas Armadas ESPE, directed by Dr. Wilbert G. Aguilar.
52
W. G. Aguilar et al.
References 1. Bellutta, P., Manduchi, R., Matthies, L., Owens, K., Rankin, A.: Terrain Perception for DEMO III. Retrieved from https://ieeexplore.ieee.org/abstract/document/898363 (2000) 2. Aguilar, W.G., Morales, S.G.: 3D environment mapping using the Kinect V2 and path planning based on RRT algorithms. Electronics 5(4), 70 (2016) 3. Eynard, D., Vasseur, P., Demonceaux, C., Frémont, V.: UAV Altitude Estimation by Mixed Stereoscopic Vision. Retrieved from https://ieeexplore.ieee.org/abstract/document/5652254 (2010) 4. Aguilar, W.G., Angulo, C.: Real-time model-based video stabilization for microaerial vehicles. Neural Process. Lett. 43(2), 459–477 (2016) 5. Fong, T., Thorpe, C.: Vehicle Teleoperation Interfaces. Retrieved from https://link.springer. com/article/10.1023/A:1011295826834 (2001) 6. Aguilar, W.G., Cobeña, B., Rodriguez, G., Salcedo, V.S., Collaguazo, B.: SVM and RGB-D sensor based gesture recognition for UAV control. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 713–719 (2018) 7. Krückel, K., Nolden, F., Ferrein, A., Scholl, I.: Intuitive Visual Teleoperation for UGVs Using Free-Look Augmented Reality Displays. Retrieved from https://ieeexplore.ieee.org/abstract/ document/7139809 (2015) 8. Aguilar, W.G., Luna, M.A., Ruiz, H., Moya, J.F., Luna, M.P., Abad, V., Parra, H.: Statistical abnormal crowd behavior detection and simulation for real-time applications. In: International Conference on Intelligent Robotics and Applications, pp. 671–682 (2017) 9. Lee, J.P.: Future Unmanned System Design for Reliable Military Operations. Retrieved from http://article.nadiapub.com/IJCA/vol5_no3/13.pdf (2012) 10. Orbea, D., Moposita, J., Aguilar, W.G., Paredes, M., León, G., Jara-Olmedo, A.: Math model of UAV multi rotor prototype with fixed wing aerodynamic structure for a flight simulator. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 199–211 (2017) 11. Matthies, L., Rankin, A.: Negative Obstacle Detection by Thermal Signature. Retrieved from https://ieeexplore.ieee.org/abstract/document/1250744 (2003) 12. Aguilar, W.G., Casaliglla, V.P., Polit, J.L.: Obstacle avoidance based-visual navigation for micro aerial vehicles. Electronics 6(1), 10 (2017) 13. Mitchelle, W., Staniforth, A., Scott, I.: Analysis of Ackermann Steering Geometry. Retrieved from https://www.sae.org/publications/technical-papers/content/2006-01-3638/ (2006) 14. Aguilar, W.G., Salcedo, V.S., Sandoval, D.S., Cobeña, B.: Developing of a video-based model for UAV autonomous navigation. In: Latin American Workshop on Computational Neuroscience, pp. 94–105 (2017) 15. Peynot, T., Underwood, J., Scheding, S.: Towards Reliable Perception for Unmanned Ground Vehicles in Challenging Conditions. Retrieved from https://ieeexplore.ieee.org/ abstract/document/5354484 (2009) 16. Aguilar, W.G., Manosalvas, J.F., Guillén, J.A., Collaguazo, B.: Robust motion estimation based on multiple monocular camera for indoor autonomous navigation of micro aerial vehicle. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 547–561 (2018) 17. Phan, C., Lui, H.: A Cooperative UAV/UGV Platform for Wildfire Detection and Fighting. Retrieved from https://ieeexplore.ieee.org/abstract/document/4675411 (2008) 18. Sathiyanarayanan, M., Azharuddin, S., Kumar, S.: Four Different Modes to Control Unmanned Ground Vehicle for Military Purpose. Retrieved from http://citeseerx.ist.psu.edu/viewdoc/ summary?doi=10.1.1.670.9990 (2014) 19. Sokolov, M., Afanasyev, I., Lavrenov, R., Sagitov, A., Sabirova, L., Magid, E.: Modelling a Crawler-Type UGV for Urban Search and Rescue in Gazebo Environment. Retrieved from https://www.researchgate.net/profile/Ilya_Afanasyev/publication/313344082_ Modelling_a_crawler-type_UGV_for_urban_search_and_rescue_in_Gazebo_environment/
Ackermann UGV with 2D Mapping for Unknown Environments
53
links/5c546cf6a6fdccd6b5da5331/Modelling-a-crawler-type-UGV-for-urban-search-andrescue-in-Gazebo-e (2017) 20. Yang, C., Jun-Jian, P., Jing, S., Lin-Lin, Z., Yan-Dong, T.: V-disparity Based UGV Obstacle Detection in Rough Outdoor Terrain. Retrieved from http://www.aas.net.cn/fileZDHXB/ journal/article/zdhxb/2010/5/PDF/100506.pdf (2010) 21. Zhao, Y., Li, J., Li, L., Zhang, M., Guo, L.: Environmental Perception and Sensor Data Fusion for Unmanned Ground Vehicle. Retrieved from https://new.hindawi.com/journals/mpe/2013/ 903951/ (2013)
Visual-Based Real-Time Detection Using Neural Networks and Micro-UAVs for Military Operations Marco Calderón, Wilbert G. Aguilar, and Darwin Merizalde
Abstract This article presents a vision-based detection system for a micro-UAV, which has been implemented in parallel to an autonomous GPS-based mission. The research seeks to determine a value objective for decision-making within military reconnaissance operations. YOLO-based algorithms have been used in real-time, providing detection of people and vehicles while fulfilling an automated navigation mission. The project was implemented in the CICTE Military Applications Research Center, as part of an automatic takeoff, navigation, detection, and landing system. The detection based on YOLO V3 offers efficient results from the analysis of sensitivity and specificity in the detection in real-time, in external environments during autonomous navigation and while the recognition of the objective is carried out keeping the UAV in stationary mode, with different angles of the camera. Keywords Object detection · Recognition · Autonomous navigation · Real-time processing
1 Introduction UAVs can recognize land or sea areas, conduct surveillance on military units, support troops through recognition, contributing to the security of strategic areas. The armies of countries with potential in military technology, everyday demand greater needs in operations for which they use unmanned aerial systems as effective support for their troops [1], causing an evolutionary change within the spatial air ordering where observation and attack practices take greater control of airspace [2]. The use of UAVs is linked to public safety, becoming an observation strategy for decision-making M. Calderón · W. G. Aguilar (B) · D. Merizalde CICTE, DEEL, Universidad de las Fuerzas Armadas ESPE, Sangolquí, Ecuador e-mail: [email protected] W. G. Aguilar FIS, Escuela Politécnica Nacional, Quito, Ecuador GREC, Universitat Politècnica de Catalunya, Barcelona, Spain © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_5
55
56
M. Calderón et al.
within the military field [3]. The Ecuadorian Armed Forces have UAVs, whose code is partially open [4, 5], which invites research to improve their capabilities based on the development of applications that allow access to their configuration and operation. Detection missions must be executed in real-time, and are subject to the adequate performance of UAV sensors and resources, mainly in the acquisition of images [4]. For the development of a UAV objective detection system, it is essential to establish a real-time communication system that allows total control from a ground station [6, 7]. At present, there is the possibility of creating applications through an SDK (Software Development Kit acronym), where the conditions for accessing UAV resources from the ground station are programmed. In the case of Parrot Bebop UAVs, the official SDK is the ARDroneSDK3. Vision-based objective detection must be established in real-time by means of algorithms integrated into intelligent autonomous systems [8, 9]. The YOLO algorithm (your only look once) is used, which uses deep learning and convolutional neural networks to classify images in real-time [4]. The detection algorithm must be implemented at the ground station, then communicate remotely with the UAV to execute the actions [10]. The use of UAVs (acronym in English of unmanned aerial vehicles) in military operations presupposes an advantage in operating within the Theater of Operations. The most powerful Armed Forces in the world are coupling UAV systems within their employment doctrine, in this context research is carried out where the use of these devices is determined in an integral way [11]. Being a technology in global expansion, the research developments in military applications are oriented to autonomous systems with sufficient capacities to fulfill all kinds of missions [12]. UAV systems have in recent years been an axis within the technological changes applied to security and defense [13]. In the Ecuadorian Armed Forces, the use of UAV systems occurs in security and surveillance operations at a tactical level [14]. All the actions carried out by UAVs are carried out by teleoperation, therefore, there are no autonomous systems implementations to fulfill specific missions within an operation. Research is currently underway for autonomous systems within the CICTE Military Applications Research Center, with excellent results, however, employment in military operations is still in process. The success of the missions entrusted to UAVs depends on the ability to perform functions in support of operations that pose a danger to people, UAVs act in their replacement with increasingly automated systems that are increasingly sophisticated. UAVs in the theater of operations are used with constant evolution, mainly in the operations of intelligence, surveillance, recognition, at the strategic operational, and tactical levels [11–14]. The contribution to military action with the use of UAVs in the theater of operations will increase when a fully automated system with route planning and specific missions assigned from a ground station is available, thus leaving the operator’s dependence on UAV control. In this way, greater precision is achieved in the mission to be accomplished, in addition it can be monitored in real-time. Currently, the use of UAVs in the Ecuadorian Armed Forces is subject to operations that help improve the state’s
Visual-Based Real-Time Detection Using Neural Networks …
57
capacity to respond to threats and internal and external risks, but through manual teleoperation. This project presents an innovative and zero-cost proposal, aimed at achieving complete automation of UAVs, optimizing the detection capabilities with which these vehicles are equipped, thus allowing the fulfillment of specific missions. Developing a vision-based detection system for MICRO-UAVs allows to fulfill missions in support of military reconnaissance operations. In the first instance, the research project proposes to establish the communication of the Micro-UAV with a ground station, replacing the manual control by a computer, in which access to the UAV resources is available. For this purpose, bebop_autonomy will be used, which is a ROS (an acronym for Robotic Operating System) controller for the Parrot Bebop drone, based on the software development kid ARDroneSDK3. During navigation towards the objective and upon reaching the exact coordinates, the detention continues using the YOLO vision-based algorithm that uses convolutional neural networks [4, 8], allowing the processing of the information provided by the on-board camera in real-time.
2 Vision-Based Objective Detection System Computer vision is the method used for target detection, in this sense a computer can detect an image based on the variation of colors, the camera becomes a sensor that provides images with a set of values to which denominate pixels and represent as a whole the dimension of photography. The representation of an image in grayscale can be observed in Fig. 1, each pixel has a value in numbers that varies proportionally according to the intensity of the color; the identification of the numerical value is done by means of a software programmed for the detection of an image in pixels [15–18]. The characteristics of the image with its representation in pixels are the first function to start with computer vision. This function is called features here different
Fig. 1 ConvNet organizes its neurons in three dimensions (width, height, depth), as visualized in one of the layers. Each layer of a ConvNet transforms the 3D input volume into a 3D output volume of neuronal activations
58
M. Calderón et al.
levels of abstraction are identified as green, colors, angles, objects of an image or characteristics of a person [19]. The features function uses tools such as shallow focus or unsupervised learning to classify images. The main focus of the detection of objectives developed is based on an unsupervised approach that works through a learning algorithm with support in Coevolutionary Neural Networks [4, 5]. It is known that YOLO is a very robust real-time image recognition system that works at high frame rates, but within it encompasses complex and thorough mathematics, so it requires a high-performance processor since images of different types are entered sizes and with a variety of multicolored combinations that within a computer are huge matrices depending on the size and therefore the processors need to perform large calculations in the shortest possible time. It is also known that before YOLO, real-time object detection systems have been designed such as RetinaNet, SSD513, among others, whose methodology is the use of classifiers and locators for detection, taking more time for processing, obtaining a detection rate of 20–25 FPS, but YOLO is different since a neural network is applied in it and this network divides the regions into bounding boxes to later analyze each frame and proceed to weigh these tables and then predict a probability percentage indicating the object that is there. The YOLO recognition system works with convolutional neural networks and its processing algorithms tend to work optimally in relation to other detection systems since they receive a single vector and it is transformed by passing through a series of hidden layers until passing through a layer of output representing class scores, as well as in YOLO its algorithmic tricks are improved. The emphasis is then placed on the comparative diagram of YOLO in relation to previously used object detection systems. The neural networks used today have been progressively developing such as the normal neural network that uses only two dimensions whose capacity supports a certain number of weights with the use of locator and classifiers that delay the calculation process if necessary. A real-time detection. In a convolutional neural network it can be seen that in its architecture it uses three dimensions which measure the same dimensions as the normal width and height, to this the third one that is depth is added so it gives us an activation volume and generating a greater amount of weights in relation to normal networks. Here we present a comparison between normal and convolutional networks. When dealing with high-dimensional inputs such as images, as we saw above it is impractical to connect neurons to all neurons in the previous volume. Instead, we will connect each neuron to only a local region of the input volume. The spatial extent of this connectivity is a hyperparameter called the receptive field of the neuron (equivalently this is the filter size). The extent of the connectivity along the depth axis is always equal to the depth of the input volume. It is important to emphasize again this asymmetry in how we treat the spatial dimensions (width and height) and the depth dimension: The connections are local in space (along width and height), but always full along the entire depth of the input volume. The YOLO prediction system uses a four-dimensional vector tx , t y , tw , th , where the x-y subscripts represent the general position of the coordinates in the image while
Visual-Based Real-Time Detection Using Neural Networks …
59
Fig. 2 Boundary tables with previous dimensions and location prediction. We predict the width and height of the frame as compensations for the centroids of the cluster. We predict the central coordinates of the frame in relation to the location of the filter application using a sigmoid function [20]
w-h represent the size and position of the bounding boxes vectors of the object to be identified are indicated [20]. If the cell is displaced from the upper left corner of the image by C x , C y and the previous bounding box has width and height pw and ph , this prediction is determinate for these equations [20]: bx = σ (tx ) + C x
(1)
by = σ ty + C y
(2)
bw = pw etw
(3)
bh = ph eth
(4)
YOLOv3 predicts an objectivity score for each bounding box using logistic regression. It should be 1 if the previous bounding box overlaps an object of fundamental truth more than any previous bounding box, we will show a diagram where the prediction method is observed using bounding boxes [20] (Fig. 2).
3 Tests and Results of the Objective Detection System 3.1 Metrics Assessed on Detection The metrics evaluated in the detection of the objects are oriented towards the ability of the system to identify people and vehicles, obtaining sensitivity and specificity. Sensitivity considers two variables, True Positive people or vehicles detected by the system and false-negative people or vehicles not detected by the system, then:
60
M. Calderón et al.
Fig. 3 Detection system during takeoff
Sensibility =
VP VP + FN
(5)
where VP True Positive and FN False Negative. Likewise, the specificity is analyzed, which is the ability to identify certain objects that are neither people nor vehicles, that is, it is considered True Negative to objects not identified as vehicles or people but that are actually present in the image. On the other hand, False Positives are those objects incorrectly identified as persons or vehicles. Specificity =
VN VN + FP
(6)
where VN True Negative and FP False Positive. The identification is usually done when a 50% probability of the object is a person or vehicle is exceeded, in which case the algorithm generates a frame around the object with the given name, as seen in Fig. 3. The tests were carried out at a height of 30 m, a distance that allows Bebop 2 to make a clear detection without putting at risk.
3.2 Description of the Tests Performed The algorithm has been programmed to detect vehicles throughout the trajectory from takeoff to mission return. Initially, the tests focus on takeoff until reaching 30 m high, during this trajectory the camera remains at 45° and the detection system is active. Under the same conditions of height and angle of the camera, detection is performed during GPS-based navigation. Finally, upon reaching the coordinate from which the target should be detected, the third test is performed by taking 10 images processed with the detection system.
Visual-Based Real-Time Detection Using Neural Networks …
61
Table 1 Sensitivity and specificity during takeoff Frames F1
VP
FN
Sensibilidad
Frames
VN
FP
7
1
0.88
F1
0
0
Especificidad –
F2
9
2
0.82
F2
2
1
0.67
F3
11
0
1.00
F3
4
2
0.67
F4
4
1
0.80
F4
3
3
0.50
F5
12
2
0.86
F5
4
1
0.80
F6
10
1
0.91
F6
2
3
0.40
F7
11
0
1.00
F7
1
0
1.00
F8
10
2
0.83
F8
3
1
0.75
F9
9
0
1.00
F9
3
0
1.00
F 10
13
1
0.93
F 10
2
2
0.50
Media
96
10
0.91
Media
53
34
0.70
3.2.1
Detection During Takeoff
There is an image at 10 m high, with a camera angle of 45°, in the building of the ESPE research center, in Fig. 3 vehicles are detected, a vehicle not detected by the system is observed. In the tests carried out during the execution of the takeoff, the data of True Positive of the image are 7, False Negative 1, True Negative 0, and False Positive 0, with this the percentage of sensitivity and specificity of the system is obtained (Table 1). Sensitivity reaches 91% and specificity 70%. It is concluded that the system works correctly in the detection of vehicles at a low height.
3.2.2
Detection During Navigation
During autonomous navigation at 30 m high, the system detects in real-time with a camera angle at 45°. In Fig. 4, vehicles are detected as True Positive and 2 false negatives, the detection of 1 Positive False, and 1 True Negative is observed (Table 2). In tests during GPS navigation, sensitivity decreases to 57% and specificity to 56%. It is concluded that the moving system has a 34% decrease in sensitivity with respect to takeoff, in terms of specificity it decreases to a lesser extent with 14%.
3.2.3
System Object Detection
For the detection of the final objective the UAV remains at 30 m high in a stationary position after the GPS navigation is completed, it takes 10 frames with a camera angle of 45°.
62
M. Calderón et al.
Fig. 4 Target detection system
Table 2 Sensitivity and specificity during GPS navigation Frames
VP
FN
Sensibilidad
Frames
VN
FP
F1
3
F2
4
2
0.60
3
0.57
F3 F4
5
2
3
2
F5
4
F6 F7
F1
1
1
0.50
F2
2
0
1.00
0.71
F3
1
2
0.33
0.60
F4
2
2
0.50
4
0.50
F5
2
1
0.67
5
4
0.56
F6
1
2
0.33
6
5
0.55
F7
2
1
0.67
F8
5
4
0.56
F8
2
2
0.50
F9
4
3
0.57
F9
1
2
0.33
F 10
3
3
0.50
F 10
3
1
0.75
42
32
0.57
Media
53
34
0.56
Media
Especificidad
In Fig. 5, 17 vehicles are detected as True Positive and 10 false negatives, the detection of 3 False Positive and 1 True Negative is observed (Table 3). In the objective detection tests, the sensitivity reaches 71% and the specificity 71%. It is concluded that the system works correctly in the detection of vehicles, Fig. 5 Target detection system
Visual-Based Real-Time Detection Using Neural Networks …
63
Table 3 Sensitivity and specificity during target detection Frames
VP
FN
Sensibilidad
Frames
VN
FP
F1
17
10
0.63
F1
3
1
0.75
F2
15
6
0.71
F2
4
0
1.00
F3
21
11
0.66
F3
5
1
0.83
F4
14
4
0.78
F4
2
2
0.50
F5
12
4
0.75
F5
3
1
0.75
F6
24
13
0.65
F6
3
2
0.60
F7
18
5
0.78
F7
2
1
0.67
F8
22
6
0.79
F8
4
3
0.57
F9
21
8
0.72
F9
5
2
0.71
F 10
16
7
0.70
F 10
3
1
0.75
180
74
0.71
Media
53
34
0.71
Media
Especificidad
the sensitivity decreases with respect to takeoff by 20%, however, with respect to the detection during navigation it rises by 14%. The specificity regarding takeoff is maintained, but it improves 15% with respect to the result during GPS navigation.
4 Conclusions and Future Works The real-time detection, parallel to the execution of a totally autonomous mission of a UAV, is possible thanks to the integration with takeoff, navigation, and landing algorithms. Convolutional neural networks with training coupled with vehicle detection. In the military application, this system is very useful because it can obtain necessary information for the operation in real-time and shortest possible time.
References 1. Orbea, D., Moposita, J., Aguilar, W.G., Paredes, M., León, G., Jara-Olmedo, A.: Math model of UAV multi rotor prototype with fixed wing aerodynamic structure for a flight simulator. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 199–211 (2017) 2. Kleinschmidt, J.: Drones y el orden legal internacional. Tecnología, estrategia y largas cadenas de acción (2015) 3. Oviedo, J.: Uso de los drones en la seguridad privada (2016) 4. Aguilar, W.G., Angulo, C.: Real-time model-based video stabilization for microaerial vehicles. Neural Process. Lett. 43(2), 459–477 (2016)
64
M. Calderón et al.
5. Aguilar, W.G., Salcedo, V.S., Sandoval, D.S., Cobeña, B.: Developing of a video-based model for UAV autonomous navigation. In: Latin American Workshop on Computational Neuroscience, pp. 94–105 (2017) 6. Aguilar, W.G., Cobeña, B., Rodriguez, G., Salcedo, V.S., Collaguazo, B.: SVM and RGB-D sensor based gesture recognition for UAV control. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 713–719 (2018) 7. Ritchie, M., Fioranelli, F., Borrion, H., Griffiths, H.: Classification of loaded/unloaded microdrones using multistatic radar. Electron. Lett. [Internet] 51(22), 1813–1815. Available from https://digital-library.theiet.org/content/journals/10.1049/el.2015.3038 (2015). Cited 2020 Jan 2 8. Aguilar, W.G., Casaliglla, V.P., Polit, J.L.: Obstacle avoidance based-visual navigation for micro aerial vehicles. Electronics 6(1), 10 (2017) 9. Vetrella, A.: On GF-2016 I 2nd IF, 2016 undefined. Cooperative UAV navigation under nominal GPS coverage and in GPS-challenging environments. ieeexplore.ieee.org [Internet]. Available from https://ieeexplore.ieee.org/abstract/document/7740606/. Cited 2020 Jan 5 10. Aguilar, W.G., Morales, S.G.: 3D environment mapping using the Kinect V2 and path planning based on RRT algorithms. Electronics 5(4), 70 (2016) 11. Campanelli, H.: La utilización conjunta de los Sistemas Aéreos no Tripulados en el Teatro de Operaciones [Internet]. Available from http://190.12.101.91:80/jspui/handle/1847939/138 (2014). Cited 2019 Dec 21 12. Ortuño, N., María, J., Cuquerella, J.L.: Desarrollo de un sistema de control para UAV con capacidad ATOL entre lanchas de instrucción de la Escuela Naval Militar. Available from http://calderon.cud.uvigo.es/handle/123456789/124 (2016). Cited 2019 Dec 21 13. Riola, J.M., Rodríguez, M.: Tecnologías del Siglo XXI TECHNOLOGIES OF THE XXI CENTURY, vol. 4 (2018) 14. Baquero Montoya, P., Briones, R.V.: EMPLEO DE LOS UAV, EN OPERACIONES DE SEGURIDAD Y VIGILANCIA EN LAS ÁREAS ESTRATÉGICAS EN EL ECUADOR, vol. IV [Internet]. Available from www.altiuas.com (2019). Cited 2019 Dec 22 15. Aguilar, W.G., Luna, M.A., Ruiz, H., Moya, J.F., Luna, M.P., Abad, V., Parra, H.: Statistical abnormal crowd behavior detection and simulation for real-time applications. In: International Conference on Intelligent Robotics and Applications, pp. 671–682 (2017) 16. Aguilar, W.G., Manosalvas, J.F., Guillén, J.A., Collaguazo, B.: Robust motion estimation based on multiple monocular camera for indoor autonomous navigation of micro aerial vehicle. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 547–561 (2018) 17. Planells, P., Blasco, F.: DESARROLLO DEL MODELO DINÁMICO DE UN CUATRIRROTOR Y DISEÑO DE LOS SISTEMAS DE CONTROL DE ESTABILIZACIÓN Y SEGUIMIENTO AUTÓNOMO DE TRAYECTORIAS [Internet]. Universidad Politécnica de Valencia. Available from https://riunet.upv.es/bitstream/handle/10251/52791/TFG_ 14340178461876400573981960261481.pdf?sequence=2 (2015). Cited 2020 Jan 10 18. Aguilar, W.G., Angulo, C.: Real-time video stabilization without phantom movements for micro aerial vehicles. EURASIP J. Image Video Process. 2014(1), 46 (2014) 19. Pusiol, P.D.: Redes Convolucionales en Comprension de Escenas [Internet]. Available from http://creativecommons.org/licenses/by-sa/2.5/ar/ (2014). Cited 2020 Jan 12 20. Redmon, J., Divvala, S., Girshick, R., Farhadi, A.: You Only Look Once: Unified, Real-Time Object Detection [Internet]. Available from http://pjreddie.com/yolo/ (2016). Cited 2020 Jan 13
Visual and Inertial Data-Based Virtual Localization for Urban Combat Wilbert G. Aguilar, Marco Calderón, Darwin Merizalde, Fabricio Amaguaña, and Jonathan Tituaña
Abstract The present work of investigation presents a system of estimation of position and orientation based on algorithms of artificial vision and inertial data taken from the unit of inertial measurement incorporated in a smartphone device. The implemented system realizes the estimation of position and orientation in real time. An application was developed for android operating systems that allows capturing the images of the environment and executes the algorithms of artificial vision. In the implementation of the system, the detectors of feature points were tested, Harris, Shi-Tomasi, FAST and SIFT, with the objective of finding the detector that allows to have an optimized system so that it can be executed by the processor of a system embedded as are smartphones. To calculate the displacement of the camera adhered to a mobile agent, the optical flow method was implemented. Additionally, gyroscope data incorporated in the smartphone was used to estimate the orientation of the agent. The system incorporates a simulation of estimated movement within a three-dimensional environment that runs on a computer. The position and orientation data are sent from the smartphone to the computer wirelessly through a Wi-Fi connection. The three-dimensional environment is a digital version of the central block of the Universidad de la Fuerzas Armadas ESPE where the tests of the implemented system were carried out. Keywords Military strategy · Augmented reality · Warlike simulator · Harris · Shi-Tomasi · Optical flow · Android
W. G. Aguilar (B) · M. Calderón · D. Merizalde · F. Amaguaña · J. Tituaña CICTE, DEEL, Universidad de Las Fuerzas Armadas ESPE, Sangolquí, Ecuador e-mail: [email protected] W. G. Aguilar FIS, Escuela Politécnica Nacional, Quito, Ecuador GREC, Universitat Politècnica de Catalunya, Barcelona, Spain © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_6
65
66
W. G. Aguilar et al.
1 Introduction The present research project focuses on the estimation of egomotion in internal environments [1], GPS denied [2], for locating military personnel during urban combat operations using monocular images and inertial information. The system will be developed in the programming environment Unity 3D and executed on the operating system for mobile phones, android. The acquisition of images will be done with a monocular camera [3], incorporated in the mobile device (smartphone). Additionally, inertial data provided by the IMU of the device will be used in order to improve the accuracy in locating the military personnel. The location of the military elements will be visualized in a virtual environment, which will be a digital version of a controlled environment. This will be executed in a computer designed to monitor military personnel. The criminal acts are not carried out only in the jungle area as it has been in the last decade; now they have moved to an urbanized zone with terrorist acts. These events have motivated the participation of military personnel in operations to support the National Police to safeguard citizen security, and it is in this field where urban combat operations become important. These combat operations are aimed at neutralizing hostile enemies, rescuing hostages, deactivating bombs, but above all, guaranteeing peace and tranquility of citizens. They are carried out in internal environments that hinder the monitoring of military personnel by the different commands and control levels. The construction of an egomotion estimation system for military units that allows monitoring military units during combat operations is proposed. This project seeks to be useful for military personnel in operations in urban environments, in which knowing the exact location is essential, since taking into account recent events in our country, and the invasion of radical subversive groups, operations of rescue or surveillance can be carried out with a geolocation backup where the trajectories of the military units and troops could be monitored, to monitor their progress or eventual extraction of them.
2 Related Works Urban combat operations are actions carried out by military forces in an environment different from conventional scenarios (mountain, jungle, air, sea), which are carried out in environments with the presence of goods, vehicles, people. These environments are called “urban.” The elements of the urban distribution must be part of the defense scheme of the zone so that the attacker can be neutralized and the effectiveness of the weapons improved. However, in these operations, military actions are affected by the morphology of urban buildings complicating the mobility of military personnel [4]. The location of military personnel is an important factor, within these operations, that can be affected by the characteristics of the environment.
Visual and Inertial Data-Based Virtual Localization …
67
The latter is due to the fact that the most common positioning systems based on satellites (GPS or Galileo) [5, 6] or on mobile telephony networks [7] are suitable for outdoor environments where there is direct vision of satellites or of the base stations. However, when these systems are used inside buildings, they present many problems with the attenuation of the received signals and the multipath caused by the rebounds of the signal. In recent years, several alternatives have been developed for locating targets in internal environments or denied GPS. One of them is the estimation of the egomotion based on images obtained by the capture device from a first-person perspective, which allows the calculation of the pose (position and rotation) of the capture device. This estimation method can be done with the use of monocular cameras, stereo cameras, RGBD cameras or 3D LIDAR sensors [8]. In the UK, an egomotion estimation system was developed by using a monocular camera installed in a vehicle. In that system, through methods such as feature matching and feature tracking, the speed of the mobile was estimated. Although the instantaneous velocity values were noisy and imprecise, the average velocity values were close to the real values demonstrating the system’s potential for estimating egomotion [9]. Another system on motion estimation of a mobile vehicle was developed in 2006. It uses a monocular camera, and through the method known as optical flow, it is possible to estimate the movement of the vehicle [10]. A favorable point for the development of systems and applications based on computer vision is the technological advance that smartphones have had with high processing capabilities in conjunction with multiple sensors. An example of this is the game called Mozzies [11] developed by Siemens being the first application to use the video camera as a sensor to create an augmented reality environment. Parallel to the development of smartphones as processing devices, software development has also taken important steps. Since 2000, the free license computer vision library has been developed. Its main advantage lies in that being multiplatform allows the development of applications in different programming softwares such as Phyton, C++, MATLAB, Unity 3D, among others [12]. Taking this background into account, this research project proposes the development of an application for mobile devices based on computer vision algorithms in order to perform the estimation of egomotion for military purposes during urban combat operations.
3 Our Approach The system comprises two stages. The first stage is called visual-inertial system and runs on a smartphone. It is responsible for capturing the images of the environment and calculating the movement parameters. The second stage is the monitoring system. This is executed on a computer. It allows the visualization of the path, estimated in the first stage, on a building plan. It also allows to observe a simulation of movement in a virtual environment.
68
W. G. Aguilar et al.
Fig. 1 Path estimation
3.1 Visual-Inertial System The visual-inertial system was implemented in a smartphone device with android operating system. The most relevant feature of the device is the Snapdragon 845 processor (2.8 GhHz) with an Adreno 630 GPU on which the algorithm developed was executed, obtaining better results than in the processors with lower clock speeds. Figure 1 shows the path estimation process. The process of estimation of position and orientation begins with the capture of the images with a rate of 30 frames per second (FPS) that is the standard quantity in the video devices.
3.1.1
Feature Points Detection
The feature points detection is based on the extraction of points from the initial image, and then, the search of said points is made in the images subsequently captured. This method is used for small movements [13–15]. In the detection of points what is sought is that they are easily distinguishable and comparable There are two types of detectors, corner detectors and texture detectors. Shi-Tomasi [16] developed a method that allows to verify the quality of the feature points between two consecutive images; for this, they use a measure of dissimilarity. If the dissimilarity between a pair of points grows a lot, the point is discarded. The OpenCV library integrates the goodFeaturesToTrack() function based on the principle of dissimilarity developed by Shi-Tomasi for the selection of features. The detection algorithm implemented in the function is a modification of the Harris corner detector. It uses an altered version of the Harris detector scoring function that determines whether or not the tracking window contains a corner [17, 18]. R = min(λ1 , λ2 )
(1)
where λ1 , λ2 are the eigenvalues of the matrix formed by the function of weights and the derivatives of the image in the directions of x and y. The values for which a
Visual and Inertial Data-Based Virtual Localization …
69
characteristic is considered a corner are given when λ1 λ2 > λmin _min where λmin is the characteristic quality threshold between [0, 1]. Algorithm 1 shows the implementation process of the Shi-Tomasi points of interest detector.
Algorithm 1: Shi-Tomasi 1: Parameters of the descriptor 2: Image captured and converted to gray scale 3: Feature points detection and description 4: Feature points stored in a matrix of points (corners) 5: The detected points plotted to observe the functioning of the function: Imgproc.circle (mask, P, 7, green); mask is the mask on which the point P is plotted, 7 is the radius of the point, and green is the color in RGB code (0,255.0) 6: The process is repeated from Step 2.
3.1.2
Optical Flow
After detecting the characteristic points, the next step is the calculation of the optical flow between two consecutive images in order to find the point correspondences and then be able to estimate the transformation matrix that allows determining the movement of the person. The technique consists of calculating the vector field where each vector is a displacement vector that shows the movement of the characteristic points between two consecutive frames. These displacement fields are calculated locally in the environment of a characteristic point. Figure 2 shows the relationship that exists in the displacement vector of an object in the real world and the optical flow vector in the image plane. That is why, the field of optical flow is defined as a projection of the movement of an object in space. Equation 2 mathematically defines the optical flow method that relates the partials of the image and velocity with respect to the x-axis and y-axis. ∂I ∂I ∂I vx + vy + vt = 0 ∂x ∂y ∂t Fig. 2 Relationship between 3D movement and movement in the plane of the image
(2)
70
W. G. Aguilar et al.
where vx = ∂∂tx and v y = ∂∂ty are the components of the velocity (optical flow) and Ix = ∂∂ xI , I y = ∂∂ yI , It = ∂∂tI are the partial derivatives of the intensity of the pixels of the image in the region surrounding a characteristic point. Restructuring
Ix , I y · Vx , Vy = −It ∇ I · vˇ = −It
(3)
where ∇ I · vˇ = −It is known as 2D movement and vˇ is the optical flow of the surrounding region one corner. OpenCV implements the calcOpticalFlowPyrLK() function that solves the optical flow equations together with the Lukas–Kanade solver to find the optical flow vectors that describe the movement of a neighborhood where a corner has been detected. Then, these vectors are compared with the vectors corresponding to the neighborhood of the corner that is in the next captured image to determine the correspondence of the pixels. These correspondences are stored in a state matrix, forming pairs of corners or points with their respective coordinates in pixels.
3.1.3
RANSAC
The set of correspondences calculated by the methods of matching, tracking or optical flow, have errors in the association called anomalous values (outliers) due to image noise, occlusion, blurring, point and vision changes and lighting. For the estimation to be more precise, it is convenient to eliminate these values by means of the RANSAC method [19, 20]; Algorithm 2 explains the process of the method.
Algorithm 2: RANSAC 1: Select the minimum number of points, useful for generating the parameter model. 2: Resolution of the parameters in the model. 3: Defines the number of points in the set that fit according to the defined tolerance. 4: If the number of points that fit over the total number of points in the set exceeds the threshold, recalculate the parameters of the model with the points that fit. 5: Otherwise, repeat Steps 1 to 4, N times.
3.1.4
Calculation of the Essential Matrix and Extraction of Movement Parameters
The estimation of visual movement consists of calculating the essential matrix (E) between two images captured in two consecutive instants of time that we will call k and k − 1. The essential matrix describes the correspondence between the points
Visual and Inertial Data-Based Virtual Localization …
71
of an image at time k − 1 and the consecutive frame k. E contains the spin data contained in the sub-matrix R and the translation data contained in the matrix t as shown in Eq. 4. E = tR
(4)
OpenCV implements the findEssentialMat() method to find the essential matrix. Algorithm 3 explains the calculation process of the transformation matrix.
Algorithm 3: Calculation of the transformation matrix 1: The first image is captured I k−1 2: High contrast points are detected by a characteristic detector, and the coordinates of the points are stored in a point matrix P. 3: The second image is captured I k 4: High contrast points are detected by a characteristic detector, and the coordinates of the points are stored in a point matrix C. 5: The optical flow between the images is calculated I k−1 , I k . 6 The correspondences between the pairs of points P i and C i are stored in a state matrix st. 7: Through the set of points C, P and the matrix of correspondences, the essential matrix E is calculated 8: The pose, rotation R and translation t that form the transformation matrix T k are recovered. 9: The image I k is copied to the image I k−1 , and the points C in the matrix of points P and Steps 3–9 are repeated
Luego de calcular la matriz esencial que contiene los parámetros de movimiento el paso final es extraerlos para ello se utilizó la función de OpenCV recoverPose() que permite descomponer la matriz esencial en las matrices R y t.
4 Results and Discussion This test was performed at 4 pm with an average illumination level of 70.57 lx. The path introduces two rotations and a considerable distance in order to pro-bar the accuracy of the system. FAST Detector: The low illumination affects this method in a remarkable way; as seen in Fig. 3, the estimated path is far from the real path. A percentage error of 35.6% was obtained. Figure 3 shows the estimated path on the smartphone screen. Harris Detector: Despite the low illumination, Fig. 4 shows only a slight difference between the two trajectories. A percentage error of 9.7% was obtained. Figure 4 shows the estimated path on the smartphone screen. Shi-Thomas Detector: As in the previous tests, the path estimated by this method turns out to be the most efficient. A percentage error of 1.3% was obtained. Figure 5 shows that there is hardly a slight difference between the two trajectories. Relative and percentage error: Table 1 shows the results obtained in the third test.
72
Fig. 3 Detector FAST—map 2D and smartphone
Fig. 4 Harris detector—map 2D and smartphone
Fig. 5 Shi-Thomas detector—map 2D and smartphone
W. G. Aguilar et al.
Visual and Inertial Data-Based Virtual Localization …
73
Table 1 Results of the path Detector
Pixels
Estimated distance
Relative error
Error (%)
Lux
70,6
FAST
244,0
19.06
0,356
35,61
Pixels
379,0
Harris
415,7
32.466
0,097
9,68
Dist
29,6
Shi-Tomasi
384,0
29.99
0,013
1,33
On this occasion, the FAST detector is presented as the most deficient with 35.6%. In the same way, as in the previous test, SHI-TOMASI is the most efficient with a percentage error of 1.3%.
5 Conclusions and Future Work The results show that corner detectors are more efficient when it comes to real-time applications, because they require less computational resources and their execution time is less. Speaking on time, Harris and Shi-Tomasi detectors allow accurate estimated trajectories to be obtained even when the ambient lighting levels are low. In the experiments carried out, four characteristic point detectors were tested; among them, the well-known SIFT detector within the field of visual odometry was tested for its robustness and precision, however not much used in real-time applications because it requires high levels of processing. With SIFT, percentage errors of 65% were obtained with differences of up to 15 m with respect to the actual distance traveled. When the levels of illumination fall considerably, in the experiments, a variation of 557 lx was registered at 70 lx; the FAST algorithm showed deficiencies in the estimation of the path with a difference of 10 m with respect to the real one. However, the Harris and Shi-Tomasi detectors showed differences of 3 m for the Harris case and barely 30 cm for the Shi-Tomasi detector. In this way, it is concluded that the ideal detector for the implemented system is the Shi-Tomasi corner detector due to its precision and also allows to have an optimized algorithm for an embedded system such as a smartphone. The algorithms of visual odometry allow to estimate the translation and rotation of the camera attached to a mobile agent.
References 1. Aguilar, W.G., Luna, M.A., Ruiz, H., Moya, J.F., Luna, M.P., Abadm V., Parra, H.: Statistical abnormal crowd behavior detection and simulation for real-time applications. In: International Conference on Intelligent Robotics and Applications, pp. 671–682 (2017) 2. Hofmann-Wellenhof, B., Lichtenegger, H., Collins, J.: Global Positioning System : Theory and Practice. Springer, Berlin (2001)
74
W. G. Aguilar et al.
3. ¿Por qué no funciona el GPS dentro de un edificio?/ Pretexsa.com. Available: http://www. pretexsa.com/RMRv9D2M.html 4. Yamaguchi, K.: Vehicle ego-motion estimation and moving object detection using a monocular camera, pp. 18–21 (2006) 5. Bevilacqua, M., Tsourdos, A., Starr, A.: Egomotion estimation for monocular camera visual odometer 6. Hannuksela, J., Barnard, M., Sangi, P., Heikkilä, J.: Camera-based motion recognition for mobile interaction. In: ISRN signal process, vol. 2011, Article ID 425621 (2011) 7. Ágila, J., Link, B., Smith, P., Wehrle, K.: Indoor navigation on wheels (and without) using smartphones (2012) 8. Nist, D., Bergen, J.: Visual odometry (2004) 9. Doctoral, T.: Un Sistema de Odometría Visual Monocular bajo Restricciones de Tiempo Real (2015) 10. Aguilar, W.G., Angulo, C.: Real-time model-based video stabilization for microaerial vehicles. Neural Process. Lett. 43(2), 459–477 (2016) 11. Aguilar, W.G., Morales, S.G.: 3D environment mapping using the Kinect V2 and path planning based on RRT algorithms. Electronics 5(4), 70 (2016) 12. Aguilar, W.G., Salcedo, V.S., Sandoval, D.S., Cobeña, B.: Developing of a video-based model for UAV autonomous navigation. Latin American Workshop on Computational Neuroscience, pp. 94–105 (2017) 13. Aguilar, W.G., Casaliglla, V.P., Polit, J.L.: Obstacle avoidance based-visual navigation for micro aerial vehicles. Electronics 6(1), 10 (2017) 14. Aguilar, W.G., Cobeña, B., Rodriguez, G., Salcedo, V.S., Collaguazo, B.: SVM and RGB-D sensor based gesture recognition for UAV control. In: International conference on augmented reality, virtual reality and computer graphics, pp. 713–719 (2018) 15. Aguilar, W.G., Manosalvas, J.F., Guillén, J.A., Collaguazo, B.: Robust motion estimation based on multiple monocular camera for indoor autonomous navigation of micro aerial vehicle. In: International conference on augmented reality, virtual reality and computer graphics, pp. 547– 561 (2018) 16. Fraundorfer, B.F., Scaramuzza, D.: Visual odometry (2012) 17. Aguilar, W.G., Angulo, C.: Real-time video stabilization without phantom movements for micro aerial vehicles. EURASIP J. Image Video Process. 2014(1), 46 (2014) 18. Szeliski, R.: Computer vision: algorithms and applications (2010) 19. Orbea, D., Moposita, J., Aguilar, W.G., Paredes, M., León, G., Jara-Olmedo, A.: Math model of UAV multi rotor prototype with fixed wing aerodynamic structure for a flight simulator. In: International conference on augmented reality, virtual reality and computer graphics, pp. 199– 211 (2017) 20. Nist, D.: An efficient solution to the five-point relative pose problem 3. The five-point algorithm
Kinect and Manipulator-Based Sample Collection System for Military Robot Orlando Caiza, Wilbert G. Aguilar, Pablo Albán, and Yéssica Fernández
Abstract The present work corresponds to the development of a sample recollection system for the exteriors robotic platform belonging to the manufacturing laboratory. The main objective of the project is the implementation of a manipulator capable of solving an automated process using artificial vision provided by the Kinect device, which provides the necessary parameters such as, required deepness for object location, all this by using free software. Through the construction of an anthropomorphic arm with six degrees of freedom which was designed for the future integration into the project realized by Obando and Sánchez which is mobile platform, an embedded system for object recollection was obtained. Additionally, a graphic interface was implemented which allows the user to visualize the recognized objects giving the user the possibility to control the object recollection operation on a controlled environment, the objects to be taken must be located within a predetermine work area. Keywords Object manipulation · Artificial vision · Anthropomorphic robot dynamics · Kinematic decoupling
1 Introduction The reason for the development of a sample collection system integrated to the robotic outdoor platform is written. The problem is analyzed giving importance to the O. Caiza · W. G. Aguilar (B) · P. Albán CICTE, DEEL, Universidad de Las Fuerzas Armadas ESPE, Sangolquí 171103, Ecuador e-mail: [email protected] W. G. Aguilar FIS, Escuela Politécnica Nacional, Quito, Ecuador GREC, Universitat Politècnica de Catalunya, Barcelona, Spain Y. Fernández Universidad Nacional de Educación, Azogues 030105, Ecuador © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_7
75
76
O. Caiza et al.
respective resolution. For this, the study of the conditions of the existing platform is planned and proceeded to condition it according to the dimensions of the manipulator robot, define the objectives to be met and determine the scope of the project.
2 Design of an Object Manipulator Robot for the Robotic Platform The Quality Matrix is developed for collecting information about the needs and requirements of the system operator. According to the procedure described on [1], the characteristics of the system to be developed are shown in Table 1. Subsystems capable of fulfilling one or more defined technical characteristics are specified in Table 2.
Table 1 QFD matrix results M
T. C.
P.
1
Frame and manipulator material
241
2
Terminal with object manipulator
219
3
Reduced cost
200
4
Robots geometry
5
Robust
M
T. C.
P.
6
Modular system
161
7
Degrees of freedom
152
8
Industrial security
122
192
9
Free software
111
191
10
Interface designed according to GEDIS guide
108
T.C. technical characteristics; P punctuation, M milestone
Table 2 Subsystems defined for design No.
S.
F.
M.
No.
S.
F.
M.
1
Structure of the manipulator robot
Orientation and position (object manipulator)
1-7
4
Robot positioning algorithm
Referential position of the robot
7
2
Power
Power system and actuators
1, 2, 3, 6
5
Artificial vision algorithm
Shapes recognition and objects positioning
4, 8
3
Control
Control system, commands and data sending
3, 8, 9
6
Human–machine interface
Interaction with the user
8, 10
Kinect and Manipulator-Based Sample Collection …
77
Table 3 Concept approach No.
Concept
Characteristics
3
• • • • •
Vision is independent of robot’s structure Moderately light structure Modular architecture pieces Low loading capability Reduced number of pieces
According to [2], the process to design a product presents two stages: • Preliminary design: The best design options are exposed, evaluated and chosen. • Final design: The designs of the selected concepts are explained. On the preliminary phase, several designs were contemplated; the one shown in Table 3 was selected. The main aspects considered for the posterior design and development of the structure are shown on Table 4. The modular architecture will allow the robot’s movements to achieve the six degrees of freedom. According to the analysis done, we selected a structural square profile with 20 mm of section length and 2 mm thickness. It is shown in Fig. 1. Now, with the selected profile, the safety factor is recalculated: n = S · σ A36 · Mmax−1 = 700 · 250 · 10, 2900−1 = 1.7
(1)
Table 5 summarizes the static analysis of the robot’s elements. Being a priority to keep the arm’s weight at minimum, it selected PLA as the construction material for the manipulator robot’s structure, and it selected aluminum for the longest slots.
Table 4 Manipulator robot’s structure required parameters Parameter
Characteristic
Arm configuration
The anthropomorphic configuration is selected because it allows to obtain better orientation and complex movements with greater ease
Load capacity
The expected load capacity is 500 g maximum
Degrees of freedom
Six degrees of freedom are necessary; three of them define the position and the remaining three orient the robot’s manipulator
Speed
The speed reached by the end effector would approach to 0.8 m/s. The opening time of the manipulator will be set to 12 s
Maximum reach
The distance required to reach an object at the base of the platform refers to a maximum of 75 cm
78
O. Caiza et al.
Fig. 1 Manipulator robot’s base (a). Selected profile for the robot’s base characteristics (b) [3]
Table 5 Static analysis of the robot’s elements Element
Part
Static analysis results
Motor’s base
Element
Part
Static analysis results
σ VM : 2.55 MPa ε: 1.03E−03 δ: 4.9E−01 mm
Base
σ VM : 5.67 MPa ε: 1.675E−01 δ: 2.55E−01 mm
Actuator’s base
σ VM : 0.563 MPa ε: 0.493E−03 δ: 3.096E−03 mm
Base axis
σ VM : 4.45 kPa ε: 0.01055E−06 δ: 7.063E−03 mm
Link’s holder
σ VM : 1.08 MPa ε: 0.286E−03 δ: 45.9E−03 mm
Actuator’s holder
σ VM : 1.58 MPa ε: 5.578E−06 δ: 5.024E−03 mm
Forces on axis analysis. Shearing force and flector moment were analyzed; forces diagram on the motor axis of the base is shown in Fig. 2. According to the procedures shown, tangential (Fy), radial (Fx) and reaction forces on the axis (Ay, By, Az, Bz) are:
F y F x Az Bz Ay By
= 2.2 1.6 2.69 2.2 1.96 1.6
(2)
Analysis to reduce the diameter. The analysis is performed on the shoulder of the axis, due to the stress concentration. On Fig. 3 is shown the tangential force at 12 mm and the radial force also at 12 mm. The resultant moment at the point of stress concentration is: Ma =
0.01762 + 0.01282 = 0.0128 Nm
(3)
Kinect and Manipulator-Based Sample Collection …
79
Fig. 2 Forces diagram on the motor axis of the base
Fig. 3 Tangential and radial forces at 12 mm
According to [4], the first iteration diameter, employing a safety factor or 2.5 and ED-Goodman criterion, is: d = 0.011 m = 11.44 mm A standard size of 20 mm is selected. The von Mises stressis: 0.035 MPa σa = σm 0.28 MPa
(4)
(5)
Considering the new diameter size of 20 mm and the recalculation of Marin size factor, the safety factor obtained using ED-Goodman criterion is: n = 515.54 This value is enough for designing the axis.
(6)
80
O. Caiza et al.
Fig. 4 Rollover analysis
Table 6 Elements that conform the final effector
Piece
Model
Piece
Gripper’s base
Slot 2
Gripper’s base support
Gripper’s transmission coupling
Gripper’s coupling
Slot 3
Slot 1
Pincer
Model
On Fig. 4 are drawn the maximum forces (F) and the position components (dx, dy) of the CG with respect to the joint. A standard safety factor of 1.5 was used. The minimum weight for avoiding rollover must be 3 kg, and the weight of the real structure is 4 kg; so, it is enough for not rollover. A sub-actuated mechanism is selected; it can handle high load objects. Its opening–closing gripper’s time is about 15 s. It is economic and has less than 20 pieces. In Table 6, the elements that conform the final effector are shown. The final effector once assembled will look like the one shown in Fig. 5. PLA is used for the most complex parts and acrylic for the simpler slots. Fig. 5 Final effector of the manipulator robot
Kinect and Manipulator-Based Sample Collection …
81
The general methodology for robots of n degrees of freedom described in [5] was employed. It gives the anthropomorphic robot parameters shown in Table 7. The following simulation was performed (Figs. 6 and 7). The dynamic behavior of the robot increases the torque’s value when it reaches a specific position. For the shoulder joint, shown in Fig. 7 (medium), maximum torque is needed when the angle oscillates on high values respect to the vertical. For the robot’s elbow, it does not require high torque, but, for low angles, with respect to the vertical, the torque tends to maximize, Fig. 7 (Right). Considering the kinematic effects over the grip force required by the pincer described in [6] and following the formulae to find the torque when employing square screws described in [7, 8], the following parameters are obtained employing a safety factor of 1.5, Table 8. According to the parameters required, the following actuators were selected (Table 9). The requirements for the microcontroller are shown on Table 10. The most suitable development platforms are: Arduino, Raspberry and ODROID because of their low cost, free software requirement and availability on the market. Data transmission media requirements are shown in Table 11. The chosen microcontroller was Arduino because it satisfies the requirements, Table 10, and it possesses analogic capacity which Raspberry and ODROID do not. Serial communication was chosen because it makes possible and effective to control the robot [9]. The sensor to be employed is Kinect; it allows to create a closed-loop system by using the object position visualization which permits to do the system main function, object recollection. It is necessary to solve the object’s position in space using the cameras and sensors available on the Kinect dispositive [10]. The main purpose of the algorithm is the identification of objects by colors. The recognized colors are green and yellow; the operations that are done to convert RGB to HSV are described in [11, 12], and the result of the algorithm for the artificial vision control system is shown in Fig. 8. For obtaining the angles of the joints and positioning the actuator, Denavit–Hartenberg method and Inverse Cinematic are used. The fixed arguments of the manipulator robot are shown in Table 12. Articulation angles were obtained using the algorithm shown in Fig. 9 [13–16]. Python with the numpy and rospy libraries is used for programming on ROS environment. D-H algorithm function is repeated the number of times necessary, so the robot can reach the objective position.
3 Test and Results Results are shown on Tables 13, 14 and 15. Video results are provided on https://www.youtube.com/watch?v=iE5gajDm830.
0.48
Nms
m
I3
0.05
kg m2
Unit
Parameter
Value
Unit
Nms
0.48
b2
kg
0.62
ms1
Nm
*
fc1
kg
1.77
ms2
Nm
1.734
fc2
kg
1.77
ms3
m/s2
9.81
g
m
0.14
cs2
Nm
Nm
15
τj2
τj1 18
kg m2
Nm
15
τj3
0.05 kg m2
I2 0.0075
I1
m
0.28
cs3
s lot; j joint; l length; m mass; I moment of inertia; b friction coefficient; fc coulomb coefficient; g gravity acceleration; τ torque; c mass center *7.17 if dθ1/dt > 0 and 8.05 if dθ1/dt < 0
b1
m
0.52
0.34
Value
ls2
ls1
Parameter
Table 7 Anthropomorphic robot parameters
82 O. Caiza et al.
Kinect and Manipulator-Based Sample Collection …
83
Fig. 6 Block diagram of the model
Fig. 7 Theta 1 response of the model at T 1 = 28 N m (left). Theta 2 response of the model at T 2 = 16 N m (medium). Theta 3 response of the model at T 3 = 12 N m (right)
Table 8 Summary of needed parameters for the actuators Motor on elbow articulation (rotation) Actuator
Torque (N m)
Speed (rpm)
Power (W)
Torque (N m)
Speed (rpm)
Gripper’s motor
0.17
110
9.4
Wrist motor
0.00945
353.706
Axis motor (motor 2)
2.53
23.873
Power (W)
Motor 3
0.192
196.906
3.97
0.35
Motor 4
12
41.062
20.91
6.33
Motor 5
16
31.322
52.63
Motor 6
0.0128
32.322
0.421
4 Conclusions Research allowed the development of an object recollection system for integration with the exterior robotic platform that belongs to the DECEM’s manufacturing Lab. An anthropomorphic robot with six degrees of freedom was designed and built, and
84
O. Caiza et al.
Table 9 Technical characteristic of the required and selected motors Motor
Power (W)
Motor1
0.421
Torque (N m)
Speed (rpm)
Selected motor (actuator)
0.0128
35.3
Motor RDS3115 ASMC 03B
Motor2
52.63
16
24
Motor3
20.91
12
19
ASMC 03B
Motor4
3.97
8.96
10
Tower Pro MG959 Servo
Motor5
6.3
18.13
6.4
RDS 3115
Motor6
0.35
10.07
19.6
Tower Pro MG959 Servo
Gripper’s motor
9
0.17
110
Tower Pro MG959 Servo
Table 10 Microcontroller requirements Requirements
Description
Requirements
Description
SPI and USB interface
For communication
At least 7 PWM outputs
For servomotors control
At least two analog inputs
For sensor lecture
Minimum clock speed
16 MHz
Table 11 Data transmission media Serial communication
Bluetooth communication
Communication type
Synchronic/A synchronic
Synchronic/A synchronic
Transmission speed
115 Mb/s
24 Mb/s
Reach (max)
1200 m
30 m
Serial communication
Bluetooth communication
Connection type
Wired
Wireless
Maximum buffer
640 bits
343 bits
it can recollect samples even in a working area below the reference level of its base. The artificial vision algorithm facilitates recognition of space position using specific colors perception, such as yellow and green; the parameters for the proper positioning of the manipulator robot also depend on environment light conditions. Positioning algorithm provides the embedded system the ability to locate effector’s end in the position obtained from the artificial vision system. Tests results are: repetitiveness of 80%, positioning error 2%, and maximum load capacity is 0.5 kg meeting the requirement raised by the user.
Kinect and Manipulator-Based Sample Collection …
85
Fig. 8 Artificial vision control system algorithm
Table 12 Fixed arguments of the manipulator robot Slot
θi
di
ai
αi
Slot
θi
di
ai
1
θ1 + 0.5π
0.18 m
0
+0.5π
4
(θ4)
0.39 m
0
2
(θ2)
0
0.38
0
5
(θ5)
0
0
3
(θ3)
0
0
+0.5π
6
(θ6)
0.26 m
0
86
O. Caiza et al.
Fig. 9 Object recollecting robot positioning algorithm Table 13 Lifting capacity tests Weight (kg)
Distance (cm)
Accomplish
Weight (kg)
Distance (cm)
Accomplish
0.1
70
YES
0.4
70
YES
0.2
60
YES
0.5
65
YES
0.3
50
YES
0.6
60
NO
Table 14 Repeatability tests Objective (cm)
Try
Time (s)
Variation (mm)
Objective (cm)
Try
Time (s)
Variation (mm)
Point 1
1
6
0
Point 2
1
7
0
2
6
1
2
7
1
3
6
2
3
7
2
4
6
1
4
7
1
Kinect and Manipulator-Based Sample Collection …
87
Table 15 Repeatability errors on point 1 and 2 Try
E. Value (cm)
M. Value (cm)
Abs. error (%)
Rel. error (%)
Point 1
Point 2
Point 1
Point 1
Point 2
Point 2
1
40
40
40
0
0
0
0
2
40
40.1
40.2
0.1
0.2
0.25
0.5
3
40
40.2
40.2
0.2
0.2
0.5
0.5
4
40
40.1
40.3
0.1
0.3
0.25
0.74
M measured; E expected
References 1. Caiza Collaguazo, E.O.: Desarrollo de un sistema de recolección de muestras para la plataforma robótica de exteriores perteneciente al Laboratorio de Manufactura del DECEM. Universidad de las Fuerzas Armadas ESPE, Matriz Sangolquí (2018) 2. Rod, F.: Montaje y Puesta en Marcahca de Sistemas Robóticos y sistemas de visión. IC Editorial, Mexico (2014) 3. IPAC: Square structural pipe (2014) 4. Budynas RG, Nisbet, J.K.: Diseño en Ingeniería Mecánica de Shigley. McGraw Hill, México (2008) 5. Ollero Baturone, A.: Robótica Manipuladores y Robots Móviles. MARCOMBO S.A, Barcelona (2001) 6. Wolf, A., Steinman, R., Schunck, H.: Grippers in motion: the fascination of automated handling tasks. Springer, Berlín (2005) 7. Mott, R.L.: Diseño de Elementos de Máquinas. Pearson Education, México (2010) 8. Romeo Ortega, A.L.: Passivity_based control of Euler-Lagrange systems. Springer, London (1998) 9. Orbea, D., Moposita, J., Aguilar, W.G., Paredes, M., León, G., Jara-Olmedo, A.: Math model of UAV multi rotor prototype with fixed wing aerodynamic structure for a flight simulator. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 199–211 (2017) 10. Aguilar, W.G., Cobeña, B., Rodriguez, G., Salcedo, V.S., Collaguazo, B.: SVM and RGB-D sensor based gesture recognition for UAV control. In: International conference on augmented reality, virtual reality and computer graphics, pp. 713–719 (2018) 11. Aguilar, W.G., Angulo, C.: Real-time model-based video stabilization for microaerial vehicles. Neural Process. Lett. 43(2), 459–477 (2016) 12. Aguilar, W.G., Luna, M.A., Ruiz, H., Moya, J.F., Luna, M.P., Abad, V., Parra, H.: Statistical abnormal crowd behavior detection and simulation for real-time applications. In: International Conference on Intelligent Robotics and Applications, 671–682 (2017) 13. Aguilar, W.G., Morales, S.G.: 3D environment mapping using the Kinect V2 and path planning based on RRT algorithms. Electronics 5(4), 70 (2016) 14. Aguilar, W.G., Manosalvas, J.F., Guillén, J.A., Collaguazo, B.: Robust motion estimation based on multiple monocular camera for indoor autonomous navigation of micro aerial vehicle. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 547–561 (2018) 15. Aguilar, W.G., Casaliglla, V.P., Polit, J.L.: Obstacle avoidance based-visual navigation for micro aerial vehicles. Electronics 6(1), 10 (2017) 16. Aguilar, W.G., Salcedo, V.S., Sandoval, D.S., Cobeña, B.: Developing of a video-based model for UAV autonomous navigation. Latin American Workshop on Computational Neuroscience, pp 94–105 (2017)
Autonomous Navigation Based on Proportional Controller with GPS Setpoint for UAV in External Environments Darwin Merizalde, Wilbert G. Aguilar, and Marco Calderón
Abstract In this paper, we propose the implementation of an autonomous navigation system for quadcopter UAVs through a trajectory established by GPS as a setpoint, where the control system, in this case, we will use the proportional one, will guide the UAV through said plotted trajectory, achieving the objective precisely, which is a very useful application in military operations since a system of this nature is required to carry out observation and recognition missions for decision making and to have clear knowledge of the scenario that is lived. To control a trajectory, the system communicates by means of a computer, acquires data by means of GPS on board and transfers them to the PC to be processed by means of calculations of conversion of distance and angle systems using global positioning data and from there, sends orders and flight plans by means of a proportional controller that allows corrections to be made with tolerances established by the controller to follow your flight plan and reach the exact point. Once reached the point, the system is validated for its automatic return, following the path of return drawn and then concludes with the landing that will be validated by artificial vision; even though the GPS system has a certain degree of accuracy, it is required to recognize a base and that you make your descent at the base where you started your flight plan, through the UAV camera vision, keeping in the system an image pattern that when identified by a certain number of points, will fulfill a landing action, ending the mission. Keywords UAV · Autonomous · Flight · Navigation · Landing
D. Merizalde · W. G. Aguilar (B) · M. Calderón CICTE, DEEL, Universidad de Las Fuerzas Armadas ESPE, Sangolquí, Ecuador e-mail: [email protected] W. G. Aguilar FIS, Escuela Politécnica Nacional, Quito, Ecuador GREC, Universitat Politècnica de Catalunya, Barcelona, Spain © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_8
89
90
D. Merizalde et al.
1 Introduction Thanks to the advancement of technology and the commercial distribution of UAVs, the use of drones has been evidenced, which is why in this document, we will discuss the importance of certain latest technology devices such as UAVs, technically named micro-unmanned aerial vehicles (UAVs) [1, 2], which have gained space in the field of research and have become very useful resources to fulfill certain specific missions in any field and environment, whether civil or military [3, 4]. At the moment, it has been seen that many people manipulate UAV by remote control, and it is incredible to observe the flexibility of movement and the ease to enter sectors that are out of the reach of any user, allowing you to better visualize and analyze your environment through an onboard camera [5, 6]. It is important to mention that it is not only human manipulation that makes this type of resource essential, but also the use of this medium autonomously. This is the essence of this investigation since it is necessary for UAVs to carry out missions autonomously only with the sending of a single order so that later the device fulfills the task and performs itself according to the order issued, using its stability controllers to avoid and correct the errors during the journey, so that later it returns with all the requested information landing safely to the exit point [5, 6]. In an operation or task to be carried out through UAV, there is a very important factor that is the mission and constitutes the essence and the reason why these technological resources are used, thus leading to the reach of the objectives required for those who need certain information that UAVs can provide [7, 8]. It is necessary to emphasize that the fulfillment of some mission or task through the use of UAVs is carried out in order to have a reach to certain areas where it is impossible for the human being to have domain of space, in this case, airspace, since that to have aerial visualization, it has always been necessary to have mobile resources of difficult physical and economic reach. For UAV operators, they need to have some training for the proper management of these, becoming a long-term investment to achieve the correct domain and operation to start it and fulfill a specific mission. This technology of micro-UAVs has allowed the human being to have access to visualize the entire environment that surrounds us from an aerial perspective and thus know everything that surrounds us and obtain data to process them according to needs [5, 6]. To give a possible solution to the problem raised, it is proposed to carry out an autonomous navigation system capable of following any trajectory set by a GPS route by means of an appropriate control action for this, as for the altitude parameter, it can be preset by the user so that the problem of obstacles does not exist during its route; a return will be applied that takes as a base the vector 0 where the UAV began its flight as a starting point, so that at that point, it makes a landing by artificial vision considering a base image which will be recognized by the camera device [5, 6]. At the time of landing, it is necessary to do it through artificial vision since the accuracy of the GPS systems is not very optimal to find the starting point, the vision system is based on the recognition of a target, and it can be. This is the symbol of a helipad or an emblem already pre-designed and loaded into the executing program [9, 10]. The
Autonomous Navigation Based on Proportional Controller …
91
recognition is based on the identification of points according to a pattern established in the digital image where the contrast and abrupt change of colors will be recorded, marking such a dimming, making a comparison and establishing a minimum of points so that when exceeding This threshold will give way to a specific task; in this case, it is to land the UAV; otherwise, it will continue in the search for the desired pattern [9, 10].
2 Related Works to the Subject This document emphasizes the autonomy of the UAV in external environments; since for interiors, it requires more precision, adjustments and programming methods for its best performance [11, 12]. For this, UAV requires very high precision and very fast location to flight in any environment to complete any mission, through a flight methodology that follows certain guidelines and steps ordered for proper execution [1, 5, 11]. For the mission, in this case, it is necessary to have a GPS module with which you can have a setpoint to follow a path in the space where you need to fly. As for the landing, it can be seen that there are different methodologies to process and execute this action, and the most option is the artificial vision that collects data through its onboard camera and allows obtaining necessary information for the process of executing an order. Successfully executing the orders and actions of the UAV will depend on the type of programming and the control action that the user can use, having enough freedom to occupy the tools that have been developed over time. Computer vision is the method used for target detection; in this sense, a computer can detect an image based on the variation of colors, and the camera becomes a sensor that provides images with a set of values which denominate pixels and represent as a whole the dimension of photography. The representation of an image in grayscale can be seen in the following figure; each pixel has a value in numbers that varies proportionally according to the intensity of the color, and the identification of the numerical value is done by means of a software programmed for the detection of an image in pixels [13, 14].
3 Our Flight Navigation Method The flight methodology for the fulfillment of the mission is established in four important phases which we will mention: Link and Communication, Takeoff and Navigation, Homework and Return Landing. The first is related to the pairing of the unit with the PC which, through the use of its connection resources, allows the connection and communication between devices that, through software and firmware, allow the understanding between the hardware to fulfill the tasks designated by the operator.
92
D. Merizalde et al.
Fig. 1 Process of communication ground air
To have a clearer idea of the system, in Fig. 1, we proceed to show the ground air link with the UAV [15–18]. The second phase is the most important since it will largely allow the development of the mission since reaching the objective corresponds to 50% of it; for this, the UAV will use GPS as a spatial orientation since it allows access to the data of global positioning that can be managed from the computer. The controller is a very important part of this phase; since for track monitoring, it is necessary to maintain a global line that will act as a setpoint for the controller whatever it is to follow the path drawn by the operator since certain disturbances such as wind, the movement of the earth, among others would change the positioning data if this did not have an optimal controller. The fulfillment of the task is essential because it constitutes in itself the reason for automating the UAV, since it specifies what the UAV is going to do and why it should reach the marked point; for this, it must be programmed in an adequate manner respecting the rules of use for professional purposes or service for humanity. As for the last phase, it constitutes the final part of the entire mission, since it is an uncomplicated task, but of great weight in the requirement, since it is very important that the unit returns in optimal conditions and with the information that it contains to be extracted and used in future operations; the fundamental part of the landing is that the UAV recognizes its site, and usually, it must be the same from where it took off, so that by means of object detection, it can land safely in any established field. Once these four important phases have been mentioned, it can be defined as a complete autonomous navigation system, capable of fulfilling any type of mission entrusted and being able to receive several missions one after another, always taking
Autonomous Navigation Based on Proportional Controller …
93
into account certain limitations in UAVs such as they can be the energy factor of the batteries or the required fuel. Making a comparison between the two modes of UAV management, autonomous flight would be more convenient since it has greater precision and greater orientation in the field to be used, since in a flight controlled by an operator on the ground, he cannot know its exact position in the airspace since it would be difficult for you to orient your UAV from a ground base with a manual control, instead with a computer-controlled flight it is possible to obtain real-time data and thus send it to the machine in real time so control your flight direction, orientation and positioning in space, thus achieving an almost immediate response when having deviations due to accusations of the natural disturbances of the environment, thus obtaining a safe navigation. Each movement is controlled according to the rotation of the propellers, their direction and speed, which is given by supplying current to the motors that support the propellers, thus giving rise and fall vertical movements, as well as pitching that allows to move laterally left and right and wink function to establish the frontal and backward movements of linear form. The combination of these motors will allow the propellers to work alternately, allowing turning movements on their own axis, clockwise and counterclockwise. The angular variables are: φ which means roll represents the swing angle, and this rotates around the x-axis, allowing the lateral, left and right displacements of the quadcopter; the symbol θ which means pitch is representing the pitch angle that revolves around the y-axis, and this angle will allow the front and backward movement of the quadcopter, and finally, the symbol ψ which means yaw represents the angle of rotation and revolves around the z-axis whose angle determines the 360° direction of the quadcopter. With all this information, the motion matrices of the Bebop 2 can already be obtained based on its dynamic reference frame with respect to its inertial reference frame; by matching the reference frames, we can determine the matrices with this equation. R CO = R(x, φ) ∗ R(y, θ ) ∗ R(z, y)
(1)
Replacing all the matrices, finally, a general matrix is obtained that will represent the movement through space. ⎡
⎤ cos y cos θ cos y sin θ sin φ − sin y cos φ cos y sin θ cos φ + sin y sin φ R CO = ⎣ sin y cos θ sin y sin θ sin φ − cos y cos φ sin y sin θ cos φ + cos y sin φ ⎦ − sin θ cos θ sin φ cos θ cos φ (2) Once the UAV rotation matrix has been obtained, this matrix can be divided into two subsystems that determine: The first is the position of the quadcopter at any point in the three-dimensional space, and the second is the orientation of the quadcopter in which it will determine the UAV course, always taking into account the UAV reference frame with respect to its inertial reference frame. Taking into account the
94 Table 1 Transfer functions of each axis of the Bebop 2
D. Merizalde et al. Axis
Transfer function
Height
G(s) =
Pitching
G(s) =
1 m∗s 2 l1 l x ∗s 2
Rolling
G(s) =
l2 l y ∗s 2
Yaw
G(s) =
1 l z ∗s 2
principle of the mass system without damping, the transfer functions can be obtained for each of the UAV axes, thus being able to control any system through its input. The transfer functions are as mentioned in Table 1. With these transfer functions, modeling can be performed, and a digital controller can be made by means of any programmer and thus determine the behavior of the system and the responses of the controller for each variable of the represented model, thus giving way to the creation of the algorithm that allows controlling the variables to precisely follow the setpoint set by the GPS system.
4 Test Scores and Discussion The proposed system has been tested and implemented in different environmental scenarios obtaining adequate and increasingly optimal results in terms of navigation accuracy, this being a reliable element to integrate a long distance mission where there is no direct vision by the human being, navigating directly to the point, allowing the user to save and optimize energy consumption for the fulfillment of any mission safely. Before presenting the navigation trajectories, the algorithm has been set up in quiet scenarios without wind and another with a certain amount of wind as proof of the disturbances that nature causes and how the system responds to it. Although the variables work together, the variables have been divided to represent them independently; in order to visualize how the controller works, first, you will obtain the graphical representations of height, latitude and length, respectively, and verify how they change depending on flight time. Here, we keep in mind the variable height where we can appreciate how it changes as a function of time (Fig. 2; Table 2). Here, we keep in mind the variable latitude where we can appreciate how it changes as a function of time (Fig. 3; Table 3). Here, we keep in mind the variable longitude where we can appreciate how it changes as a function of time (Figs. 4, 5, 6 and 7; Tables 4, 5 and 6). Allowing to determine that the controller allows a certain error threshold that is noticeable in the face of wind disturbances (Fig. 8). In the representation, it is observed that in favorable conditions, the system is of smooth movement, but when wind disturbances are present, sudden movements can be noticed in the flight; according to the analysis, the positioning correction due to
Autonomous Navigation Based on Proportional Controller …
95
Fig. 2 Comparison of variable height on stage without wind and winds
Table 2 Comparison of variable height on stage without wind and winds Scenario
Arrival time at maximum height
Time of flight
1. Windless
1 min
3 min 20 s
2. Wind moderated
1 min 19.92 s
6 min 11.628 s
Fig. 3 Comparison of variable latitude on stage without wind and winds
Table 3 Comparison of variable latitude on stage without wind and winds Scenario
Time of arrival at the point of destination
Time of arrival to the base
1. Windless
1 min 27 s
3 min 20 s
2. Wind moderated
2 min 41.57 s
6 min 11.628 s
the disturbances makes a greater consumption of battery being thus vulnerable to a possible fall by wear of energy in the case of a longer route.
96
Fig. 4 Comparison of variable longitude on stage without wind and winds
Fig. 5 Comparison of variable speed on stage without wind and winds
Fig. 6 Comparison of battery consumption on stage without wind and winds
D. Merizalde et al.
Autonomous Navigation Based on Proportional Controller …
97
Fig. 7 Comparison of trajectory in 2D latitude and longitude on stage without wind and winds Table 4 Comparison of variable longitude on stage without wind and winds Scenario
Time of arrival at the point of destination
Time of arrival to the base
1. Windless
1 min 27 s
3 min 20 s
2. Wind moderated
2 min 41.57 s
6 min 11.628 s
Table 5 Comparison of variable speed on stage without wind and winds Scenario
Average speed (km/h)
Average return speed (km/h)
1. Windless
70
70
2. Wind moderated
65
60
Table 6 Comparison of battery consumption on stage without wind and winds Scenario
Percentage when starting the plan (%)
Percentage when ending the plan (%)
Total consumption (%)
1. Windless
64
52
12
2. Wind moderated
65
46
Fig. 8 Comparison of trajectory in 3D height latitude and longitude on stage without wind and winds
98
D. Merizalde et al.
5 Conclusions and Future Works The implemented system allows to carry out a mission where human vision has no reach thanks to the fact that it has a camera on board. Based on the graphical representations of the movement variables in three-dimensional space analyzed in Sect. 4, the navigation system by a proportional controller is reliable and valid for any environment. Based on the tables in Sect. 4, it can be determined that the autonomous flight is very precise to reach the desired point taking into account the scope it should have in terms of communication with the UAV and the battery reserve that must be calculated to integrate any mission. The problem of battery consumption remains a limitation in terms of the use of UAVs, so the controller tries to design that its response to disturbances is as smooth as possible, thus avoiding sudden movements that tend to consume more amount of energy. For a future job, a better control action can be performed since it will allow smoother control of the system and thus have a better optimization of the batteries, to perform complete perimeter routes. Another work in the future can be done a graphical interface of the system that will allow monitoring the UAV and verify its position in real time, allowing to be found in case of loss or fall due to shocks and energy expenditure.
References 1. Venkatesh, G.A., Sumanth, P., Jansi, K.R.: Fully autonomous UAV. In: Proceedings: 2017 International Conference on Technical Advancements in Computers and Communications, ICTACC 2017, pp. 41–44, Oct 2017 2. Orbea, D., Moposita, J., Aguilar, W.G., Paredes, M., León, G., Jara-Olmedo, A.: Math model of UAV multi rotor prototype with fixed wing aerodynamic structure for a flight simulator. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 199–211 (2017) 3. Rathbun, D., Kragelund, S.: IEEE Xplore—an evolution based path planning algorithm for autonomous motion of a UAV through uncertain environments. Digital Avionics (2002). Available from: http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1052946%5Cnpapers2: //publication/uuid/8479FE2D-6988-48FE-B714-669DA578E2E3 4. Aguilar, W.G., Morales, S.G.: 3D environment mapping using the Kinect V2 and path planning based on RRT algorithms. Electronics 5(4), 70 (2016) 5. Sani, M.F., Karimian, G.: Automatic navigation and landing of an indoor AR. Drone quadrotor using ArUco marker and inertial sensors. In: 1st International Conference on Computer and Drone Applications: Ethical Integration of Computer and Drone Technology for Humanity Sustainability, IConDA 2017, pp. 102–107, Jan 2018 6. Aguilar, W.G., Manosalvas, J.F., Guillén, J.A., Collaguazo, B.: Robust motion estimation based on multiple monocular camera for indoor autonomous navigation of micro aerial vehicle. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 547–561 (2018) 7. Rahman, M.F.A., Radzuan, S.M., Hussain, Z., Khyasudeen, M.F., Ahmad, K.A., Ahmad, F., et al.: Performance of loiter and auto navigation for quadcopter in mission planning application using open source platform. In: Proceedings of 7th IEEE IEEE International Conference on Control System, Computing and Engineering (ICCSCE), 2017, Nov 2017. pp. 342–347 (2018)
Autonomous Navigation Based on Proportional Controller …
99
8. Aguilar, W.G., Cobeña, B., Rodriguez, G., Salcedo, V.S., Collaguazo, B.: SVM and RGB-D sensor based gesture recognition for UAV control. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 713–719 (2018) 9. Shang, J, Shi, Z.: Vision-based runway recognition for UAV autonomous landing. IJCSNS Int J Comput Sci Netw Secur 7(3), 112 (2007). Available from: http://paper.ijcsns.org/07_book/ 200703/20070317.pdf 10. Aguilar, W.G., Luna, M.A., Ruiz, H., Moya, J.F., Luna, M.P., Abad, V., Parra, H.: Statistical abnormal crowd behavior detection and simulation for real-time applications. In: International Conference on Intelligent Robotics and Applications, pp. 671–682 (2017) 11. Brockers, R., Hummenberger, M., Weiss, S., Matthies, L.: Towards autonomous navigation of miniature UAV. IEEE Comput Soc Conf Comput Vis Pattern Recognit Work 645–51 (2014) 12. Aguilar, W.G., Salcedo, V.S., Sandoval, D.S., Cobeña, B.: Developing of a video-based model for UAV autonomous navigation. Latin American Workshop on Computational Neuroscience, pp. 94–105 (2017) 13. Imanberdiyev, N., Fu, C., Kayacan, E., Chen, I.M.: Autonomous navigation of UAV by using real-time model-based reinforcement learning. In: 2016 14th International Conference on Control, Automation, Robotics and Vision (ICARCV), Nov 2016, pp. 13–5 (2017) 14. Aguilar, W.G., Casaliglla, V.P., Polit, J.L.: Obstacle avoidance based-visual navigation for micro aerial vehicles. Electronics 6(1), 10 (2017) 15. Aguilar, W.G., Angulo, C.: Real-time video stabilization without phantom movements for micro aerial vehicles. EURASIP J. Image Video Process. 2014(1), 46 (2014) 16. Aguilar, W.G., Angulo, C.: Real-time model-based video stabilization for microaerial vehicles. Neural Process. Lett. 43(2), 459–477 (2016) 17. Parrot. Dron Parrot Bebop 2 FPV|Sitio Web Official de Parrot [Internet]. 2019 [cited 2020 Jan 8]. Available from: https://www.parrot.com/es/drones/parrot-bebop-2-fpv 18. Roberts, C.: GPS guided autonomous drone, 32 (2016). Available from: https://www.evansville. edu/majors/eecs/downloads/projects2016/CameronRobertsReport.pdf
Monte Carlo-Based Localization for Kidnapped Robot Problem Wilbert G. Aguilar, Darwin Merizalde, Marco Calderón, and Alexis Carrera
Abstract In this paper, we propose an algorithm for kidnapped robot problem based on Monte Carlo localization in known map using laser 360°, and initial sample in all global localization. We present performance y real-time video results. Our proposal is evaluated and compared with other works. Keywords Kidnapped robot problem · Monte Carlo localization · Beam model sensor, UGVs
1 Introduction Unmanned ground vehicles (UGVs) have experimented an increasing interest in several research topics within localization and autonomous navigation [1]. Common applications in localization and autonomous navigation include slam [2–4], path planning [5–8], odometry [9, 10], global localization [11], and others. A common problem in autonomous localization and navigation is the kidnapped robot problem, which consists of the mobile robot being able to locate itself in its global position from any position at the moment of it is start or kidnapped. In [12], talks about algorithms used for the resolution of this problem, and using a comparative table, we find the Monte Carlo localization (MCL), which is based on a particle filter, is a robust algorithm for this problem. There are several works to solve the kidnapped robot problem using different algorithms. The best known can be found in [12] and they are EKF, MHT, coarse grid, fine grid, MCL [14, 16], where the latter is the most used by researchers and W. G. Aguilar (B) · D. Merizalde · M. Calderón · A. Carrera CICTE, DEEL, Universidad de Las Fuerzas Armadas ESPE, Sangolquí, Ecuador e-mail: [email protected] W. G. Aguilar FIS, Escuela Politécnica Nacional, Quito, Ecuador GREC, Universitat Politècnica de Catalunya, Barcelona, Spain © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_9
101
102
W. G. Aguilar et al.
therefore has some modifications such as [15], there are also works that use new algorithms such as MSAC [13]. In terms of odometry, investigations are carried out depending on the configurations of the mobile robot, such as [9, 10], or methods with sensors such as odometry using a laser sensor [17]. Our approach is resolve the kidnapped robot problem based on the Monte Carlo location (MCL), on a known map using a 360° laser with the theory of beams sensors model, and our theory about minimum sampling to start the MCL algorithm, to get a minimum error in the final location. This paper is organized as follows: our proposal for localization for kidnapped robot problem is described in Sect. 2. In Sect. 3, we present the experimental results. Finally, conclusions and future works are presented in Sect. 4.
2 Our Approach Our proposal to recover the localization of the robot in 2D is use the algorithm Monte Carlo localization (MCL) with KDL-Sampling and a 360° laser sensor with the beam model sensor, and will create samples for all the map random with a simple sampling theory to acquire the initial sample, which consists of creating 40 samples per square meter of map. Monte Carlo Localization with KDL-Sampling: For resolving kidnapped robot problem, we use Monte Carlo localization (MCL) algorithm, the basic idea is approximate the subsequent state of a set of sample states or particles xt[m] , and in a summarized way, it consists of a two-step algorithm [18]. Initialization: At time t = 0, draw M particles according to p(x0 ) = X 0 [m] Recursion: At time t > 0, generate a particle xt[m] for each particle xt−1 by drawing [m] from the actuation model p xt |u t , xt−1
= X t . Then, draw M particles from X t ,
draw with a probability proportional to measurement p(z t |xt[m] ). Call the resulting set of X t . For sampling, we use KDL-Sampling, the key idea is to bound the error introduced by the sample-based representation of the particle filter. To derive this bound, we assume that the true posterior is given by a discrete, piecewise constant distribution such as a discrete density tree or a multidimensional histogram [19, 20]. If we choose the number of samples n as [21]: n=
1 2 x 2 k−1,1−δ
(1)
A good approximation of (1) is: 3 k−1 2 2 n= + z 1−δ 1− 2 9(k − 1) 9(k − 1)
(2)
Monte Carlo-Based Localization for Kidnapped Robot Problem
103
where z1−δ is the upper 1 − δ quantile of the standard normal N(0, 1) distribution and is upper bound on the K. Combining these two analyzes we obtain the algorithm used in our proposal.
where: X t = Estimated position; u t = Control Variable z t = Measurement;n = number of samples First, we initialize all variables equal to 0, then we start the number of samples equal to the Minitial . In line 4 of the algorithm, we have the estimation of the movement of each sample, depending on the control variable, which in our case is the odometry of the robot. In line 5, the weight of the sample is analyzed, in line 6, the weight is added to the normalization factor, and it is inserted in X t (line 7). Then, the sample is entered in a discrete distribution dependent on the weight and it has an upper bound ε on the K-L distance.
104
W. G. Aguilar et al.
In line 9–11, the factor k is analyzed, and here, we assume that a bin of the multinomial distribution has support if its probability is above a certain threshold. This way the number k will decrease with the certainty of the state estimation. In line 13, the number of samples needed by the process in its next stage is calculated using the factors of the multinomial distribution, for our proposal, the maximum error between the true distribution and the estimated distribution, should not exceed 0.05. Then, on line 15–17, the weight of the sample is normalized, and the most probable estimated x positions are analyzed according to the weight of the sample and maintained and added to the X t positions, while the least probable ones are erased and they are resampling in the most probable positions. As the MCL is an algorithm that estimates the position according to the odometry of the robot, it can present high positioning errors of the robot when problem have been solved the problem, and give a false position, if it does not start with a number of acceptable samples for the map. For tests carried out in this bibliography [12], it is necessary to have more samples for greater precision, so infinite samples are required, but since it is not possible in real world, a certain number of samples must be given to start the problem to be resolved with great precision and at the same time not force the processing of the robot. Our approach for the number of initial samples in the MCL algorithm consists of creating forty samples in each square meter of known map. Minitial = 40 ∗ A
(3)
where: Minitial = Number of samples in time = 0; A = the number of square meter in known map Beam Model Sensor: This method has the central task to determine P(z|x), the probability of a measurement z given that the robot is at position x, using in step 5 of algorithm KDL-Sampling with MCL. For that, we have four equations represented a probability P(z|x) [12].
Phit (z|x, m) = n √
1 2π b
2 1 z − z exp 2b e −
Punexp (z|x, m) = nλe−λz n z max n Pmax (z|x, m) = z small Prand (z|x, m) =
(4) (5) (6) (7)
Monte Carlo-Based Localization for Kidnapped Robot Problem
⎞T ⎛ ⎞ Phit (z|x, m) αhit ⎜ αunexp ⎟ ⎜ Punexp (z|x, m) ⎟ ⎟ ⎜ ⎟ P(z|x, m) = ⎜ ⎝ αmax ⎠ ⎝ Pmax (z|x, m) ⎠ Prand (z|x, m) αrand
105
⎛
Then, the constant proposals would be the following: For Laser Scan (Beam model Sensor): Laser max beams = 100 Laser max range = 6.0 (m) αhit = 0.45 αunexp = 0.05 αmax = 0.05 αrand = 0.45 b = 0.2 λ = 0.1 For Algorithm: Initial sampling dependent on the map area: Minitial = 40 ∗ A Number of samples using KDL-Sampling z 1−δ = 0.99 = 0.05
(8)
106
W. G. Aguilar et al.
3 Results and Discussion The robot built to perform the tests has a linear velocity of 0.2 m/s, an angular velocity of 2 rad/s, which also govern the update of the algorithm, since the algorithm is updated with the movement of the robot. First, the results obtained by the tests in different places with different surfaces are presented in the following sequence of images of Fig. 1, where the time elapsed since the start of the recovery process of the kidnapped robot problem is also described; the actual position of the robot is presented, then a comparison of the results obtained from the time and the error is presented at the end of the measurement, where the calculated error is: e = |zr eal − x|
(9)
wher e z real = measurement taken from a front wall to the robot; x = Measure taken in the program with built-in measuring tool. Then, we describe the results compared with other related works; however, you cannot make a comparison of time with respect to such jobs because they depend on the speed of the robot. Error in model b: The error is presented in a map of 125 m2 in Table 1, to then compare it with errors in other related works in subjectively large maps given that it has no measurement. e = |z real − x| Totalerror position =
ex2 + e2y = 0.158 m
Comparison with related works. The error of the position in the final state of the position estimation is compared with results obtained by Dierter Fox in particle filters for mobile robot localization [22] in Table 2, where in the reference we have multi-robot and single robot compared with our approach. Video results are provided on https://www.youtube.com/watch?v=0FEfAq6JhJg.
4 Conclusions and Future Work Our proposal for the recovery of the position for kidnapped robot problem has a better estimate with respect to single robot with respect to the position error, however, with respect to the multi-robot, it has a lower estimate as shown in Table 2, but these results are subjective because the number of samples of each algorithm must also be measured and compared with the known map surface. The algorithm works correctly, because it comes to converge in an estimated final position, and thanks to the KLD-Sampling allows to reduce the number of dependent
Monte Carlo-Based Localization for Kidnapped Robot Problem
b.1)
a.1)
a.1) Algorithm test on known map of size 9x6 m => b.1) Algorithm test on known map of size 12x10 m =>
a.2)
b.2) a.2) t=5s time after the process started b.2) t= 24 s time after the process started
a.3)
a.4)
b.3) a.3) t=19 s time after the process started b.3) t=37s time after the process started
b.4) a.4) t=23s time after the process started b.1) t=49 s time after the process started
Fig. 1 Results of our approach
107
108
W. G. Aguilar et al.
Table 1 Error estimated Measure with respect to the origin of the map
Position x (m)
Position Y (m)
Angular θ (rad)
Real
0.96
4.26
6.17
Estimated
1.01
4.11
6.25
Error
0.05
0.15
0.08
Table 2 Compared with other work
Algorithm
Error position
Single robot [22]
1000 cm
Multi-robot [22]
2 cm
Algorithm KLD-Sampling with MCL
15.8 cm
samples to the estimation of the robot, with that in the end the processing is much smaller than at the beginning of the algorithm. This algorithm can be combined with the adaptive Monte Carlo localization (AMCL) which allows to generate random samples to recover the position of a possible measurement error, given that, if the robot is misplaced, generating random samples can allow finding a position with more likely than the current one. Acknowledgements This work is part of the project perception and localization system for autonomous navigation of rotor micro-aerial vehicle in GPS-denied environments, VisualNavDrone, 2016-PIC-024, from the Universidad de las Fuerzas Armadas ESPE, directed by Dr. Wilbert G. Aguilar.
References 1. Royer, E., Lhuillier, M., Dhome, M., Lavest, J.-M.: Monocular vision for mobile robot localization and autonomous navigation. Int. J. Comput. Vision 74(3), 237–260 (2007). https://doi. org/10.1007/s11263-006-0023-y 2. Savaria, D. T., Balasubramanian, R. (2010). V-SLAM: vision-based simultaneous localization and map building for an autonomous mobile robot. 2010 IEEE Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI), pp. 1–6. http://doi.org/10.1109/MFI. 2010.5604466 3. Aguilar, W.G., Casaliglla, V.P., Polit, J.L.: Obstacle avoidance based-visual navigation for micro aerial vehicles. Electronics 6(1), 10 (2017) 4. Leonard, J. J., Durrant-Whyte, H. F. (1991). Simultaneous map building and localization for an autonomous mobile robot. In: Proceedings IROS ’91: IEEE/RSJ international workshop on intelligent robots and systems ’91, (91), pp. 1442–1447. http://doi.org/10.1109/IROS.1991. 174711 5. Garcia, M.A.P., Montiel, O., Castillo, O., Sepúlveda, R., Melin, P.: Path planning for autonomous mobile robot navigation with ant colony optimization and fuzzy cost function evaluation. Appl Soft Comput J 9(3), 1102–1110 (2009). https://doi.org/10.1016/j.asoc.2009. 02.014
Monte Carlo-Based Localization for Kidnapped Robot Problem
109
6. Aguilar, W.G., Morales, S.G.: 3D environment mapping using the Kinect V2 and path planning based on RRT algorithms. Electronics 5(4), 70 (2016) 7. Ismail, A.-T., Sheta, A., Al-Weshah, M.: A mobile robot path planning using genetic algorithm in static environment. J. Comput. Sci. 4(4), 341–344 (2008). Retrieved from http://thescipub. com/abstract/10.3844/jcssp.2008.341.344 8. Ge, S.S., Cui, Y.J.: New potential functions for mobile robot path planning. IEEE Trans. Robot. Autom. 16(5), 615–620 (2000). https://doi.org/10.1109/70.880813 9. Aguilar, W.G., Cobeña, B., Rodriguez, G., Salcedo, V.S., Collaguazo, B.: SVM and RGB-D sensor based gesture recognition for UAV control. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 713–719 (2018) 10. Aguilar, W.G., Salcedo, V.S., Sandoval, D.S., Cobeña, B.: Developing of a video-based model for UAV autonomous navigation. Latin American Workshop on Computational Neuroscience, pp. 94–105 (2017) 11. Aguilar, W.G., Manosalvas, J.F., Guillén, J.A., Collaguazo, B.: Robust motion estimation based on multiple monocular camera for indoor autonomous navigation of micro aerial vehicle. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 547–561 (2018) 12. Orbea, D., Moposita, J., Aguilar, W.G., Paredes, M., León, G., Jara-Olmedo, A.: Math model of UAV multi rotor prototype with fixed wing aerodynamic structure for a flight simulator. In: International Conference on Augmented Reality, Virtual Reality and Computer Graphics, pp. 199–211 (2017) 13. Majdik, A., Popa, M., Tamas, L., Szoke, I., & Lazea, G. (2010). New approach in solving the kidnapped robot problem. In: 41st International Conference on Robotics, pp. 1–6. Retrieved from http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5756811 14. Dellaert, F., Fox, D., Burgard, W., Thrun, S.: Monte Carlo localization for mobile robots. In: Proceedings 1999 IEEE International Conference on Robotics and Automation (Cat. No. 99CH36288C), vol. 2, pp. 1322–1328 May 1999. http://doi.org/10.1109/ROBOT.1999.772544 15. Thrun, S., Fox, D., Burgard, W., Dellaert, F.: Robust Monte Carlo localization for mobile robots. Artif. Intell. 128(1–2), 99–141 (2001). https://doi.org/10.1016/S0004-3702(01)00069-8 16. Fox, D., Thrun, S., Burgard, W., Dellaert, F.: Particle filters for mobile robot localization. Sequential Monte Carlo Methods in Practice, 401–428 (2001). http://doi.org/10.1007/978-14757-3437-9_19 17. Censi, A.: An ICP variant using a point-to-line metric. In: Proceedings—IEEE international conference on robotics and automation, pp. 19–25. http://doi.org/10.1109/ROBOT.2008. 4543181 18. Aguilar, W.G., Angulo, C.: Real-time model-based video stabilization for microaerial vehicles. Neural Process. Lett. 43(2), 459–477 (2016) 19. Moore, A. W., Schneider, J., Deng, K.: Efficient locally weighted polynomial regression predictions. In: Proceedings of International Conference on Machine Learning, pp. 236–244 20. Aguilar, W.G., Luna, M.A., Ruiz, H., Moya, J.F., Luna, M.P., Abad, V., Parra, H.: Statistical abnormal crowd behavior detection and simulation for real-time applications. In: International Conference on Intelligent Robotics and Applications, pp. 671–682 (2017) 21. Fox, D.: KLD-sampling: adaptive particle filters. Adv. Neural Inf. Process. Syst. 14(1), 713–720 (2002). https://doi.org/10.1.1.21.5786 22. Fox, D., Thrun, S., Burgard, W., Dellaert, F.: Particle filters for mobile robot localization. Sequential Monte Carlo methods in practice, pp. 401–428 (2001). https://doi.org/10.1007/9781-4757-3437-9_19
Automatic Counting of People in Crowded Scenes, with Drones That Were Applied in Internal Defense Operations on October 20, 2019 in Ecuador Henry Cruz, Rolando P. Reyes Ch., and María Pinillos Abstract Eleven days of public protests, unrest and chaos happened in October 2019 in Ecuador. The national security forces acted immediately implementing internal defense operations. In this changing and highly uncertain context the automatic crowd counting based on the Gaussian density estimation combined with the multiscale convolutional neural networks is used to generate electronic intelligence. This solution is set in a real scene obtaining excellent results and reducing the training time without being the performance affected. In this way, with the information obtained in real-time about the quantities of forces and dispositive. The national security forces reaction is immediate guaranteeing the control, the resupply and reorganization of themselves. Keywords Security operations · Real scenarios · Crowd counting · CNN · Gaussian density
1 Introduction
1.1 Context On October 3, 2019 the Ecuadorian government applied the 883 decree-law which established an increase in the fossil fuels cost. This fact triggered a wave of parades, incidents and public protests carried out by different social actors. Those actions were exceptionally violent as it had not happened in previous decades [1]. H. Cruz (B) · R. P. Reyes Ch. Universidad de Fuerzas Armadas-ESPE, Sangolquí, Ecuador e-mail: [email protected] M. Pinillos Universidad Nacional de Chimborazo, Riobamba, Ecuador © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_10
111
112
H. Cruz et al.
In this context, social and political issues were evident and taken into consideration and it can be thought this situation could stem from a constant ideological fight, an economic disagreement and even manipulation of people. The official and nonofficial media broadcasted the day by day parades which took place along the country while confusing citizens witnessed it and were affected in one or another way by this officially declared state of emergency. Several social and political theories have turned up from these events pointing out the social discontent or a complex political strategy as the origin of the disruption of the proper functioning of the main cities and a try to destabilize the government legally formed. Independently of the origin of these events the destruction, confrontation and unrest were unprecedented [2]. On the other hand, the more significant social protests were those which were organized by some indigenous leaders and came from different areas of the “Interandina” region [3] from both the North and the South towards the capital of Ecuador, Quito. It is located at 2850 m above the sea level and its population is 2,700,000 inhabitants [4]. The conflict led to antagonisms and confrontations on social networks. It is where the information provided by official media was questioned by citizens due to the amount of videos, audios and fake news broadcasted on Internet leading to misleading information and uncertainty in the country. This social behavior of ideological suggestion and collective manipulation showed that the majority of the Ecuadorian society does not analyze or verify information and the fragility of the social structure that reached the social and political polarization being a consequence of people’s frustration [5]. All of these variables mentioned above intensified the public protests towards the Centre of Quito being estimated the number of participants in them about 15,000 people [6]. In this context the national armed forces accomplishing the constitutional and Ecuadorian laws work to guarantee the social peace, rights and freedom of Ecuadorian citizens. In order to do so tactics, techniques and procedures were used to have internal control. Among the technological tools applied to obtain information of the context to be confronted are the drones. They do surveillance and recognition tasks and if they are combined with computer vision algorithms, drones provide the obtaining automatic information about the composition, the device, and the effectiveness generating more accurate assessments in internal and external operations [7, 8].
1.2 People Counting in Crowded Scenes The artificial vision atomizes the surveillance and monitoring tasks through the development of algorithms that let, among other actions, the recognition, the people counting and the detection of anomalous events. Especially, they allow the people counting
Automatic Counting of People in Crowded Scenes …
113
in crowded scenes so they can be monitored making easier for those in charge of the national security the control of a situation. People counting in crowded scenes are a technique in demand in continuous development and improvement. To the best of our knowledge one of the first works in literature is related to the application of the Gaussian process regression (GPR) that consists of the segmentation of homogeneous components in a dynamic way as the background is separated from the foreground. According to the authors, this method can reach the 91% of accuracy [9]. On other hand, algorithms have been applied to the automatic learning without supervision to detect and estimate the amount of people in crowded scenes. This procedure is called density learning and allows the considerable reduction in the execution time in real time, the higher the interaction number, the lower the execution time and mistakes [10]. After the people detection and counting algorithms that do the detection through the geometric filter have been developed or designed and the subsequent people monitoring taking into account a base line detector. Whole system is based on the density and supervised training [11]. Then works related to counting people through semi supervised algorithms, accumulative attribute, extraction of spatial temporal features, multi-source and multi-scale counting and interactive object [12]. In addition, adaptive models have been developed to people counting proposes through convolutional neural networks [13], Bayesian model [14], random forest [15], density estimation by Gaussian extraction model crowded scene in perspective by deep learning [15–17]. These works have led to the presentation of proposals based on iterative crowd counting, scale aggregation network, learning multi-level density maps, pan density crowd, multipolar normalized density map, adaptive density maps, 3D crowd counting via Gaussian kernels among others [18, 19].
1.3 Drones and Crowd Counting Currently drones have become into a very useful tool to surveillance and security fields. Because of that the technological development in this area focuses on generating applications that allow the improvement of penetration, flexibility and versatility abilities that enable drones to carry out surveillance actions. Given the fact that the growth of this activity is evident it is foreseen the need of a regulation of the same [20, 21]. Technologies such as the Internet of things, artificial intelligence, artificial vision, high energetic performance, video stabilization among others improve the abilities, autonomy, design and accessibility of the drones up to a point that it would let the massive action of mobile surveillance [22]. The advantages and challenges in relation to the scope and capacity of these systems lead to doubts about the capacity to secure
114
H. Cruz et al.
and avoid the interference and intrusion in the control and monitoring systems being another challenge to be overcame by the cyber security [23]. The methods and algorithms developed for the crowd counting look for reaching higher accuracy and reducing errors. The increase in the processors capacity and in the graphic accelerators makes the detection more accurate with less execution time. These features are used by the drones to the acquisition and processing of images and the video frames via off-line and on-line [24–26]. The optical sensors of the drones produce high resolution images and videos and the stabilization systems are stronger so the likelihood of error because of bad information quality is each time smaller [27]. These abilities and advantages help obtaining more accuracy in the people detection, crowd counting and targets location and for that algorithms based on Layout Proposal Networks (LPNs) [28] and spatial kernels have been applied and the scale-aware contextual extraction to the obtaining of density maps and later training of the obtained features [29]. In relation to the crowd counting other works combine the Gaussian density estimation classified according to CNN what improves the detection even though a distortion happens due to perspective [30]. Also this combination has proved to be efficient even when imagines vary of scale due to the vertical movements of the drones [31]. Currently there is a tendency towards the use of drones for monitoring and surveillance tasks being the detection algorithms and automatic counting an important reliable option to execute these tasks. Because of that this work is to show the experiential application of the combination of the estimation of the Gaussian density with the CCN in internal defensive operations. This work has the following sections. Section 2 describes the applied methodology; Sect. 3 shows the results obtained against the application scenario are discussed and finally, Sect. 5 details the main conclusions and future work.
2 Methodology 2.1 The Application of the Gaussian Density Estimation After previous analysis of existing works the application of the density estimation through the Gaussian model applied to the people counting and improving the counting in perspective is proposed [9, 32, 33]. The beginning proposal is that this model responses to the existence of a signals distribution but which are in a defined rank where a function m(x) is directly related to its covariance k(x, x ), so it is determined that that covariance represents a vector and a matrix at the same time. The model is proposed as it is shown in (1): f ∼ GP(m, k),
(1)
Automatic Counting of People in Crowded Scenes …
115
2.2 Classification Through the CNN Signals are quantified and generate numerical matrixes which are used to develop training processes that will serve to matrixes comparison and later predictive process will be done. Given the fact that the signals taking under the same perspective facilitates the comparison what implies high levels of error because of the movement and displacement of the drones, the extraction of the training signals features in different scales lets cover higher number of possibilities being the values matching effective and subsequently the classification increases the accuracy level. Because of that the signals training through the multiscale convolutional neutral network (MSCNN) proposed by Zeng et al. [34] is thought to be an interesting option when generating a training model for the detection and crowd counting in diverse scales. The covariance model applied for the signal analysis allows having a values group and estimating the variation grade in relation to the size of the signals.
2.3 Application of the Process The process starts with the acquisition of video signals which have been treated in two ways. One of them is through an online system generated by an Ad Hoc net of the DJI Mavic 2 PRO drone to process the video in real time and another is through an off-line system processed form the DJI Phantom 4 drone. An Intel Pentium 7 with graphic accelerator computer was used to process the video and imagines. On other hand, the acquired signals by the optical sensors are assessed through the method based on MSCNN and following the process described in the item 2.2., obtaining a final map of Gaussian density. The probability of the Gaussian density is defined as the group of pixels accumulated with a certain value inside of a specific region as it is described in (2): p(x, y) =
N 1 (x − xi )2 + (y − yi )2 exp − C i=1 2ϑ
(2)
where p represents a normalised value of p from (0, 1). Figure 1 shows the process taken. The total of the video frames and the images with standard resolution of 4096 × 2160 to 60 fps and taken with a bit rate of 100 Mbps samples was of about 4000. This data bank of images was collected along the eleven days of the public protests happened in Quito in October 2019. This data bank cannot be published because of the confidentiality of the information and its repository is located at Centro de Investigación de Aplicaciones Militares (CIAM). The environmental and the application conditions varied; due to the black smoke produced by the tires burning the visibility was reduced so the drones took images from different range of heights from
116
H. Cruz et al.
Fig. 1 Counting crowd process through Gaussian density and MSCNN
20–300 m and in perspective with angles of 30 and 45 in relation to the horizon. Figure 2 shows a processed images set and the obtained results.
2.4 Response Time Optimization The defensive operations are characterized by the uncertainty and given the fact that the conditions in an urban conflict scene are changing, the response time is to be immediate. The technological solutions to support the operations in these kind of sceneries should be flexible and appropriate. The training process was optimized and taking into consideration to reach 97% of effectiveness, the training should be done in 10 periods [34]. However, take this process with conventional computers means days of processing so the training chosen counted with 30% of what is recommended, three periods. In order to reduce the error each frame was assessed in three occasions, after that the threshold was determined considering the variance of the values in the Eq. (3) N Var =
1
x−X n
2 (3)
Automatic Counting of People in Crowded Scenes …
a) Process in Ave. 12 de Octubre
117
b) Process in Ave. Tarqui
c) Procesamiento en el Arbolito Park Fig. 2 Process in different places in the Quito Centre: a Av. 12 de Octubre, b Av. Tarqui, c Arbolito Park
where x, represents a register value counting from the image 1 to the image 3. This process reduced the error due to the dispersion control and improved the response time because of the reduction in the training time for the subsequent use of the algorithm.
3 Results Assessment The comparative technique of the value counting automatic system against a Ground Truth that is based on a probabilistic counting estimating the number of people in certain amount of pixels have been applied. As every image were taken and executed by the drones in real time, the results have been stored and assessed in a laboratory in real time leading to set a reliable reference pattern in relation to the operations executed. In this way, to set or calculate the performance the mean absolute error
118 Table 1 Comparative table of results in crowd density estimation is different levels
H. Cruz et al. Level
MAE
RMSE
50
23.1
34.6
150
24.2
37.1
300
25.0
39.2
(MAE) and the root mean square error (RMSE), that are metrics traditionally used to assess accuracy of crowd counting data, have been used. The assessments have taken place in three levels, L1 = 50, L2 = 150, and L3 = 300 m, the number N = 400 samples have been taken from the CIAM data base as it is shown in Table 1. As it can be observed in Table 1 the values variations of the metric values have a propositional relation with the height level to which the frames or images were taken. It can be said that as higher is the height level, higher is the increase in the error. This variation is not taken into account due to the excellent resolution of the images obtained from the drones.
4 Discussion The multiscale model reaches high accuracy and versatility levels. However, in the training and recognition process the great amount of values produced by the covariance model demands high computational load leading to longer training time. In the context of national internal defense, the scenes are changing, and the responses should be flexible, reliable and immediate and the algorithm based on MSCNN suit to those requirements. The strategy based on the reduction in the number of training periods let gain opportuneness along the operations and the compensation mechanism to reduce error through the application of the assessment of the counting results variance led to the dispersion reduction. The crowd counting results were accurate as Table 1 shows and let have processed information which is considered electronic intelligence because generated information in relation to the forces and the equipment of the forces. Also given the geo positioning function of the drones the obtained data was geo-referenced.
Automatic Counting of People in Crowded Scenes …
119
5 Conclusions and Further Work The public protests in October 2019 in Quito led to an uncertain and chaos environment where the national security forces had to act accomplishing the constitutional law. The need to provide immediate response to guarantee the control let the implementation of the solutions to this level through the crowd counting algorithm based on the map of Gaussian density and the classification based on MSCNN. To reach a better response in implementation time with high accuracy levels, the training periods are reduced and the variance covers a certain number of obtained results so the values dispersion is reduced and adapt to a threshold. The implementation was applied in a real scene of conflict and the results were effective leading to automatized information that later on became into electronic intelligence information. Authorities are to carry on with the developed work done so as to improve the online broadcasting performance from different perspective and to apply resources of clusters servers to reduce the training data time. Acknowledgements This work was supported by Centro de Investigación de Aplicaciones Militares (CIAM) through management internal projects. Henry Cruz give thanks the support of the officers Darwin Merizalde and Marco Calderón.
References 1. Arteta, C., Lempitsky, V., Noble, J., Zisserm: Interactive object counting. In: European Conference on Computer Vision. Cham (2014) 2. Carvajal, A.: Quito se convirtió en la ciudad más poblada del Ecuador con más de 2,7 millones de habitantes en el 2018. Recuperado el 2019 de Ene. de 11, de (2019 de Ene. de 10). http:// www.elcomercio.com 3. Carvin, S.: Canadian Defence and new technologies. In: Canadian Defence Policy in Theory and Practice. Toronto (2020) 4. Chan, A., Liang, Z., Vasconcelos, N.: Privacy preserving crowd monitoring: Counting people without people models or tracking. In 2008 IEEE Conference on Computer Vision and Pattern Recognition (2008) 5. Chan, A., Liang, Z., & Vasconcelos, N.: Privacy preserving crowd monitoring: counting people without people models or tracking. In: 2008 IEEE Conference on Computer Vision and Pattern Recognition (2008) 6. Choi-Fitzpatrick, A., Juskauskas, T.: Up in the air: applying the Jacobs crowd formula to drone imagery. Procedia Eng (2015) 7. Comercio, E. (03 de Nov. de 2019). El Comercio. Obtenido de El Comercio: https://www. elcomercio.com/pages/muertos-protestas-octubre-ecuador-decreto.html
120
H. Cruz et al.
8. Cruz, H., Eckert, M., Meneses, J., Martíne: Precise real-time detection of nonforested areas with UAVs. IEEE Trans. Geosci. Remote Sens. (2016) 9. Cruz, H., Eckert, M., Meneses, J., Martínez, J.: Efficient forest fire detection index for application in unmanned aerial systems (UASs). Sensors (2016) 10. Fennelly, L., Perry, M.: Unmanned aerial vehicle (drone) usage in the 21st century. In: The Professional Protection Officer (2020) 11. Fu, M., Xu, P., Liu, Q., Ye, M., Zhu, C.: Fast crowd density estimation with convolutional neural networks. Eng. Appl. Artif. Intell. (2015) 12. Grubesic, T., & Nelson, J.: Drone futures. In: UAVs and urban spatial analysis (2020) 13. Haque, S., Sadi, M., Rafi, M., Islam, M.: Real-time crowd detection to prevent stampede. In: Proceedings of International Joint Conference on Computational Intelligence (2020) 14. Hsieh, M., Lin, Y., Hsu, W.: Drone-based object counting by spatially regularized regional proposal network. In: Proceedings of the IEEE International Conference on Computer Vision (2017) 15. Ilyas, N., Shahzad, A., Kim, K.: Convolutional-neural network-based image crowd counting: review, categorization, analysis, and performance evaluation. Sensors (2020) 16. Krishnaveni, P., Sutha, J.: Novel deep learning framework for broadcasting abnormal events obtained from surveillance applications. J. Ambient Intell. Humaniz. Comput. 1–15 (2020) 17. Küchhold, M., Simon, M., Eiselein, V., Sikora, T.: Scale-adaptive real-time crowd detection and counting for drone images. In: 2018 25th IEEE International Conference on Image Processing (ICIP) (2018) 18. Lempitsky, V., Zisserman, A.: Learning to count objects in images. In: Advances in neural information processing systems (2010) 19. Liu, B., Vasconcelos, N.: Bayesian model adaptation for crowd counts. In: Proceedings of the IEEE International Conference on Computer Vision (2015) 20. Liu, W., Lis, K., Salzmann, M., Fua, P.: Geometric and physical constraints for drone-based head plane crowd density estimatio. In: IEEE/RSJ International Conference on Intelligent Robots and Systems (2019) 21. Liu, W., Salzmann, M., Fua, P.: Context-aware crowd counting. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2019) 22. Maya, A.: La fragilidad ambiental de la cultura. Instituto de Estudios Ambientales (1995) 23. Morales, A.: Latinoamérica arde pero no brilla. Obtenido de (2019). https://es.panampost.com/ editor/2019/11/29/latinoamerica-arde-pero-no-brilla/ 24. Onoro-Rubio, D., López-Sastre, R.: Towards perspective-free object counting with deep learning. In: European Conference on Computer Vision (2016) 25. Pham, V., Kozakaya, T., Yamaguchi, O., Okada: Count forest: co-voting uncertain number of targets using random forest for crowd density estimation. In: Proceedings of the IEEE International Conference on Computer Vision (2015) 26. Rasmussen, C.: Gaussian processes in machine learning. In: Summer School on Machine Learning (2003) 27. Reyes C.R., Pérez, V.H., Paredes, M., Aguilar, W.: MilNova: an approach to the IoT solution based on model-driven engineering for the military health monitoring. CHILEAN Conference on Electrical, Electronics Engineering, Information and Communication Technologies (2017) 28. Rodriguez, M., Laptev, I., Sivic, J., Audibert: Density-aware person detection and tracking in crowds. In 2011 International Conference on Computer Vision (2011) 29. Siddappaji, B., & Akhilesh, K.: Role of cyber security in drone technology. In: Smart Technologies (2020) 30. Sindagi, V., Patel, V.: A survey of recent advances in cnn-based single image crowd counting and density estimation. Pattern Recogn. Lett. (2018)
Automatic Counting of People in Crowded Scenes …
121
31. Villalba, J.: Reporte de Desarrollo Humano en ciudades selectas del Ecuador. Recuperado el 2019 de Feb. de 13, de (2019 de Feb. de 13). http://Scribd.com 32. Vivares, E.: La “Batalla de Quito”. Análisis Carolina 27(1) (2019) 33. Von Borstel, M., Kandemir, M., Schmidt, P., Rao, M.: Gaussian process density counting from weak supervision. In: European Conference on Computer Vision (2016) 34. Zeng, L., Xu, X., Cai, B., Qiu, S., Zhang, T.: Multi-scale convolutional neural networks for crowd counting. In: 2017 IEEE International Conference on Image Processing (ICIP) (2017)
Expert Nutritional System for Military Athletes Based on Fuzzy Logic and Inferential Statistics Diana Vallejos, Freddy Tapia , Hernán Aules, Michelle Torres, and Cristian Bejarano
Abstract Athletic performance seeks to enhance the physical and mental state of the athlete, and it is implicit to systematize sports training; for this reason, managing the monitoring of the physical state of athletes implies relating variables such as initial physical performance, hydration, types of nutrition, and training to take successful strategic actions. Among the innumerable factors that promote success in sport, training, motivation and, above all, resistance to bodily injury is limited since the time to test the athlete in respect to his talent and training the margin of risk between victory and defeat is minimal, nutrition is a fundamental element in the physical preparation of a specific athlete of those who are very disciplined because the diet benefits or affects sports performance and therefore the desired result. It must be taken into account that each athlete must be aware of the nutritional objectives to be achieved and how they can improve a feeding strategy to comply with the expert’s guidance and in this way shape their long, medium and short term goals. This study was a focus on planning and training of military athletes and proposes the structure of an expert nutritional system to reach better standards. The system is available through the use of web services focused on new technologies based on inferential statistical systems and fuzzy logic applying a statistic analysis aimed at decision making. The web services are aimed at improving the monitoring of the physical condition of the D. Vallejos · F. Tapia (B) · M. Torres · C. Bejarano Department of Computer Science, Universidad de las Fuerzas Armadas-ESPE, Sangolquí, Ecuador e-mail: [email protected] D. Vallejos e-mail: [email protected] M. Torres e-mail: [email protected] C. Bejarano e-mail: [email protected] H. Aules Department of Mathematics, Universidad Central del Ecuador, Quito, Ecuador e-mail: [email protected] © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_11
123
124
D. Vallejos et al.
athletes to provide nutritional support for the maintenance of a healthy lifestyle with good physical performance, that is to say, it simplifies the information contained in the original variables, looking for the interrelations between the numerical variables defined simultaneously in a set of elements. The results achieved show an improved performance in the physical evaluation of athletes. Keywords Athletes performance · Inferential statistics · Fuzzy logic · Inference engine · Fuzzification
1 Introduction In Ecuador, for several years, the relationship between those who reach the category of high-performance athletes compared to a number of athletes who remain in the practice of a certain sport is low; despite the efforts of local and national coaches and managers [1]. Therefore, it is imperative to look for alternatives that improve this relationship; that is why, the present study addresses one of the fundamental relationships to achieve this desired improvement, the training–feeding relationship [2]. The purpose of assessing the physical condition of an athlete is to tend to continuous improvement, for which it is important to determine causes of improvement or stagnation in performance, since this information is essential to make decisions related to budgets, training programs, feeding programs, technological equipment, among others [3]. The World Health Organization recommends boosting levels of physical activity, within the Global Strategy of Physical Activity and Nutrition [2], based on today´s needs society to counter-act the rates of low activity or physical exercise. Physical activity levels can be classified and/or measured in athletes using different assessment methodologies, this classification can be manifested on ordinal scales (inactive, moderate or active), dichotomous (active or inactive) or continuous (according to METS and kilocalories) [4]. Actually is a growing interest in the study of the relationship between health and physical activity, in which a key role has been determined the relationship between the increase in cardiovascular diseases with sedentary lifestyles [5]; With this background, it is sought to link the technological aspect with the sports aspect, for which the automation of traditional processes and the optimization of the analysis times and possible actions in real-time are intended. What will allow athletes to be constantly evaluated and monitored?
Expert Nutritional System for Military Athletes Based …
125
2 Related Work Some studies highlight that sports performance is perfectly related to the technological aspect, such that a key element is a diffuse logic, which allows to represent mathematically uncertainty and vagueness. So it is that with the analysis of variables and contribution of the specialist, the levels of flexibility and muscular power are better achieved, based on better nutrition and performance. Here is the convenience of using a system that learns from the experiences and data generated and projected.
3 Methodology, Experimental Configuration, and Tools (Current References) Expert systems are considered the first truly operational product of artificial intelligence. They are computer programs designed to act as a human specialist in a particular domain or area of knowledge. In this sense, it’s a system that uses stored knowledge and some methods of inference in order to solve problems normally required by human experts, who transmits his knowledge to the system, and the user who uses it to solve a problem with the effectiveness of the specialist [1]. The expert system uses the knowledge stored and some methods of inference [3]. The fundamental feature of an expert system is that it separates the stored knowledge (knowledge base) from the program that controls it (motor of inference). Other fact important is the data of a given problem is stored in a separate database (fact-based). The characteristics of fuzzy logic are perfectly coupled for the elaboration of a nutritional system because as already mentioned above to give an effective and reliable diagnosis, closed values are not taken; otherwise, other decision criteria specific to this tool are chosen [4, 5]. Another important aspect of this system is that virtually all quantitative data collected throughout the entire process are values that are described in traditional terms, such as low, medium and high average pressure. The nutritional system uses fuzzy logic [6] developed in x-fuzzy in which input and output variables, logical operators, defuzzification with the center of the area method and Fuzzy Mean [7], and rule base were established; the system has two motors of analyzing each person’s data to give them a category and give nutrition recommendations on a case-by-case basis (Figs. 1 and 2). Another important aspect of the nutritional expert system is that it is built with PHP so that it can be used through the web, it uses an MYSQL database to save each person’s records according to a control date. With this data using a model that uses functions performed in JavaScript, the system displays statistical graphs and the progress of each person. In other words, technology is used optimally and efficiently.
126
D. Vallejos et al.
Fig. 1 Inference engine to make the category of the data of each athlete, based on the suggestions of the expert nutritionist
Fig. 2 Deployment of the diffuse system, where the input variables, data processing and output variables are detailed, resulting in the different inference engines, for the sample of results
4 Getting Data The first step in providing nutritional control is to collect personal information about each athlete, followed by an interpretation and analysis of that information to identify problems that affect the athlete’s nutritional. Among the data to be obtained are
Expert Nutritional System for Military Athletes Based …
127
Fig. 3 Fragment of the list of athletes with the display of the data of the variables taken to perform the inferential analysis, the same ones that will serve to experience the changes in the diagnosis of their physical performance
sex, age, weight, height, glucose, triglycerides, diastolic pressure, systolic pressure, temperature, and speed. To carry out this study, 80 cadets from Eloy Alfaro High School are considered, of which half are considered high-performance athletes and the other half are not (Fig. 3).
5 Data Analysis The main components method provides a series of multivariate procedures and models that reduce the size allowing better handling and interpretation of the available data by finding new variables (main components). For the corresponding analysis, we have p numeric variables x1 , x2 , x3 , . . . , x p defined simultaneously in a set of elements, and it is a question of constructing, from these variables, linear combinations y1 , y2 , y3 , …, y p uncorrelated in each other, so that the former express the largest proportion of the information contained in the values of the original variables. These linear combinations are known as major components; where the first major component y1 is expressed as follows: y1 = u 11 x1 + u 12 x2 + · · · + u 1 p x p
(1)
With the necessary condition that it captures the greatest variability of the data. The second main component y2 , must be not correlated with the first and must capture the greatest variability that has not been expressed by the first. The third main component must be not correlated with the previous ones, this as the first two must capture the greatest variability that has not been expressed by the first two, and so on.
5.1 Choosing the Number of Main Components The amount of information of the original p variables is completely reproduced by the corresponding main components. In practice, the criterion of considering a number
128
D. Vallejos et al.
of major components that retains between 70 and 95% of the total variation in the correlation between variables is used, or the Kaiser rule could be used, in which each major component should express an amount of information greater than what each variable explains.
5.2 Analysis for the Proposed Research The concepts and properties presented refer to the entire population where the variables are defined; however, they may be limited to a representative sample of it. In this way, they can be applied to the S matrix of sample covariance or to the matrix of sample correlations. If the results of the sample are to be extended to the entire population, it should be based on samples that are not small (recommended 15 cases for each original variable) the same as those to be representative of the population. Based on the data presented in the investigation and based on the following information (Fig. 4).
Fig. 4 Original variable data
Fig. 5 Correlated data
Expert Nutritional System for Military Athletes Based …
129
Fig. 6 Big plot predictive data of the main components
Based on the data presented in the research and from the information in Figs. 5 and 6, after the respective analysis is to be, the equations of the linear combinations of the 14 variables that lead to the generation of the main elements are the following: y1 = −0.19065050z 1 − 0.02521255z 2 + 0.55678264z 3 − 0.27269462z 4 + · · · + 0.4375967926z 14
(2)
y2 = −0.26341149z 1 + 0.44529228z 2 − 0.02443611z 3 + 0.10511275z 4 + · · · − 0.0094515325z 14
(3)
... y14 = −0.51543598z 1 − 0.13374224z 2 + 0.13329924z 3 + 0.13612686 + · · · − 0.6924780724z 14
(4)
On the other hand, the variables of the entire population in the proposed research are clearly observed to be well correlated and may be limited to a representative sample of it by applying a sample correlation matrix, starting from a sample of 14 variables with the observation of 20 data per variable constituting a representative sample in the research (Fig. 7). On the other hand, the trace or sum of the diagonal elements of the correlation matrix is 14.00 and equal to the sum of the variances of the original standardized
130
D. Vallejos et al.
Fig. 7 Auto valor and principals components
variables. The sum of the first self-values (3.08487047 + 1.86029536 − 4.945166) corresponds to 35.32261% of the total sum, such as the main components, at least should retain between 70 and 95% of the total correlation variation between variables, then by the time by which by sequential analogy, we will perform it at least up to the sixth self-value that corresponds to 75.45862% of the total variation, indicating that the variables are well correlated, however, it can be considered at most up to the tenth self-value that 94.14443% and on this, up to the fourteenth self-value would correspond to 5.8556% (100 − 94.14443 − 5. 8556), the same that could be omitted from the main components, because they were close to the 95% confidence limit in the field of research, with a statistically standardized 5% margin of error (Fig. 8). Analytically it can be said that all the variables initially proposed in the research are highly correlated with the variables x1 , x2 , x3 , …, x14 generated by the corresponding frequencies of the respective variances (s). In other words, the variables to be considered would be: x1 , x2 , x3 , x4 , x5 , x6 , x7 , x8 , x9 , x10 de las x1 , x2 , x3 , …,
Fig. 8 Increasing variance of the main components
Expert Nutritional System for Military Athletes Based …
131
Fig. 9 BMI analysis chart
x14 ; however, if analyzed more closely, all variables are closely correlated, meaning that variables are important and essential throughout research reaching 100% of observations.
6 Results The system uses the named Takagi-Sugeno approach, which is a method of fuzzy logic that allows to work with the data collected previously. In this method an expert has to specify his knowledge in the form of linguistic rules [1, 8]. The nutritional expert system allows us to carry out the categorization of each person by means of the rules established in the fuzzy logic and according to the registered data, also allows to recommend a diet suitable to each person for their correct nutrition with the to help improve your physical performance. Another important aspect of the system is that by means of graphs it allows us to visualize the progress corresponding to each person, so that the expert or the professional in charge can make decisions. The desired results allowed to determine the causes and consequences of good nutrition, which will be a very important contribution to the athletic rise of military personnel. [8] (Figs. 9 and 10).
7 Conclusions and Future Works Considering the parameters that motivated this research, it makes it clear that the present relates topics immersed in the profile of the professional in Physical Culture, Sport and Recreation, such as investigative skills, as well as highlighting the athletes’ responsibility to contribute to health approaching prevention and promotion of the correct nutrition. In the permanent construction of wellness alternatives to generate
132
D. Vallejos et al.
Fig. 10 Graph of the components of medical examinations
quality of life. Given that health is a priority today, which is why elite athletes and health organizations promote lifestyle habits that encourage physical activity. The expert system developed is a tool for sports initiation, but to complement it is important recommended that will research the proper exercise routine suitable for each person, which can help improve performance in the sport where it was chosen. By way of conclusion, analytically it can be said that; all the variables initially proposed in the investigation are highly correlated with the variables x1 , x2 , x3 , .., x14 generated by the corresponding frequencies of the respective variances, summarizing; the variables to be considered would be the following: x1 , x2 , x3 , x4 , x5 , x6 , x7 , x8 , x9 , x10 de las x1 , x2 , x3 , . . . , x14 , however, if analyzed more closely, all the variables are closely correlated, this means that the variables are important and essential in all the research reaching 100% of the observations, this can also be observed in (Fig. 8) the same that has a behavior of increasing functional monotony among the variables involved in the investigation.
References 1. González Rico, R., Ramírez Lecguga, J.: Review of Physical Fitness Assessment Tests in Secondary Education. Spain (2017) 2. Cajamarca Cadme, E.L., Cajamarca Cuji, H.A.: Lipid Profile in Athletes Belonging to the Azuay Sports Federation, Cuenca 2017. Cuenca (2017) 3. Hernandez, E.F.: Relationship Between Physical Exercise and Body Temperature in Physically Active and Inactive Adults. Bogota (2014) 4. Ochoa Granda, E.G.: Assessment of Initial Physical Performance and Strategic Actions to Reduce Sports Dropout in Women’s Athletics of the Sports Federation of Loja. Sangolquí (2018) 5. Cordente Martínez, C.A., García Soidán, P., Sillero Quintana, M., Domínguez Rome-ro, J.: Relationship of the level of physical activity, blood pressure and body fat in Madrid adolescents (2007) 6. Quisbert Quispe, U.: Expert System Based on Fuzzy Logic for the Child-Bearing Initiation of Children in the Growth Stage. La Paz (2015) 7. Riza c. Berkan sheldon L.trubatch.: Fuzzy System Design Principles. IEEE Press (1997) 8. Hudson, d.l., Cohen, m.e.: Neural Networks and Artificial Intelligence for Biomedical Engineering. IEEE Press series on Biomedical Engineering 1999 Wiley-IEEE Press (1999)
Expert Nutritional System for Military Athletes Based …
133
9. Véliz, C.: Multivariate Analysis, Multivariate Statistical Methods for Research. Editorial Cengage Learning, Mexico (2016) 10. Cuadras, C.: New Methods of Multivariate Analysis. CMC Editions Editorial, Barcelona (2019) 11. Montanero, J.: Multivariate Analysis. Editorial University of Extremadura, Spain (2008) 12. Baillo, A., Grané A.: 100 Solved Problems of Multivariate Statistics, implemented in Matlab. Delta Publishing Publications, Madrid (2008) 13. ALP, A.: Effects of aerobic exercise on bone-specific alkaline phosphatase and urinary CTX levels in Premenopausal Women. BMC Musculoskelet Disorders, pp. 30–31 (2013) 14. Maimoun, L., Manetta, J., Couret, I., Dupuy, A.M., Mariano-Goulart, D., Micallef, J.P., Rossi, M.: The Intensity Level of Physical Exercise and the Bone Metabolism Response, vol 27. US National Library of Medicine National Institutes of Health (2011) 15. Taha, L.G.E.-D.: The Role of Expert System in Remote Sensing Applications, vol 18. Nova Science Publishers, Inc., p. 3/4 (2010) 16. Lake, B.M., Ullman, T.D., Tenenbaum, J.B., Gershman, S.J.: Building Machines That Learn and Think Like People. Cornell University (2016) 17. Bushra Alhijawi, Y.K.: Int. J. Adv. Intell. Paradigms. Inderscience Publisher (2018) 18. Mamlook, R.M.: A New Method for Defuzzification and Ranking of Fuzzy Numbers Based on the Statistical Beta Distribution, vol. 2016 Hindawi (2016) 19. Gil-Sierra, D.G., Salcedo-Parra, O.J., Angulo, K.: Optimization using diffuse logic of analysis device for chemical components of natural ingredients based on the internet of things IoT. Sci. Mag. 1(37) (2020) 20. Bermudez, N.S.C., Tovar, S.M.C.: Analysis of the Physical Condition of the Military Cadets Academy General José María Córdova¨. Dialnet, pp. 131–139 (2015)
Earth Coverage Model for GPS-Like Capabilities Vaughn H. Standley, Edward A. Boucheron, Robert K. Kirkwood, and Benjamin E. Norman
Abstract In terms of the number of satellites in view from a near ground level device or event where Maximum Look Angle represents the performance of transmitters, the global coverage provided by the GPS constellation is demonstrated to be well approximated by assuming that the satellites are randomly distributed around the earth. Analysis using GPS almanac data indicates that a random approximation is a good approximation to within about 10% accuracy. An analytic model relying on this approximation was used to assist in the validation of a sophisticated ephemeris-based performance model. The model also enables fast parametric estimation of aggregate cost, performance, and risk of GPS-like capabilities, including those fielded on small satellites in low earth orbit. Keywords Global positioning system · Risk analysis · Software validation
1 Introduction For more than 50 years, the National Nuclear Security Administration and its predecessors have sponsored work at the National Nuclear Laboratories to develop and
V. H. Standley (B) National Defense University, Washington, DC 20319, USA e-mail: [email protected] E. A. Boucheron The Aerospace Corporation, P.O. Box 5800, Albuquerque, NM 87185, USA e-mail: [email protected] R. K. Kirkwood Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, CA 94551, USA e-mail: [email protected] B. E. Norman Los Alamos National Laboratory, P.O. Box 1663, Los Alamos, NM 87545, USA e-mail: [email protected] © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2020 Á. Rocha et al. (eds.), Developments and Advances in Defense and Security, Smart Innovation, Systems and Technologies 181, https://doi.org/10.1007/978-981-15-4875-8_12
135
136
V. H. Standley et al.
fabricate instruments to detect nuclear detonations from space. Synergies with spacebased navigation led to these instruments being hosted on GPS satellites beginning in the 1980s since both navigation and detonation detection missions require global coverage and the use of highly accurate clocks to determine location. While coverage provided by these systems is normally assessed using sophisticated orbital ephemeris-based models [1], it is possible to provide reasonable estimates with a compact analytic model. Estimates from this method can be used to help inform acquisition decisions and as a validation check of performance estimates provided by more complicated models. The method shown here calculates the probability that a specific number of satellite-based instruments will receive signals emitted from the earth’s surface or, conversely, that radio receivers on the ground will receive signals from a specific number of satellites. This method, not previously used, is based on modeling satellites’ sensing behavior with a single parameter, instrument Maximum Look Angle, and can be used to generate aggregate system performance of a complex satellite constellation.
2 Maximum Look Angle The sensitivity of satellite-based instruments to events on the surface of the earth can be represented very simply in terms of a maximum “Look Angle” that does not depend on details of local terrain and uses the approximation of a spherically symmetric earth and atmosphere. Maximum Look Angle (MLA) is the greatest angle off zenith to where satellite-based instruments are still sufficiently sensitive to detect a surface event of specified characteristic signals, such as a nuclear explosion. Look Angles are used because they consider the attenuation of propagating signals through the atmosphere. This attenuation is least when observing directly down to the earth, and greater when looking obliquely toward the earth’s horizon. Figure 1 illustrates the GPS satellite
Fig. 1 Relationship between Maximum Look Angle (MLA), Elevation Angle (θ), Satellite Cone Angle (α), and Earth Central Angle (β)
MLA
s rsat Atmosphere
Event or Device Event Horizon
re Earth Surface
Earth Coverage Model for GPS-Like Capabilities
137
relationship between the event, observing instrument, earth, and subtended angles. Maximum look angles are computed by the sensor developers based on analysis and testing. For example, the implied Maximum Look Angle associated with a hand-held GPS receiver is almost 90°, meaning that an earth-based observer can receive signals from all satellites above the earth horizon. For purposes of a simple mathematically abstracted representation, the relationship between Maximum Look Angle and the number of average satellites in view is treated as having a strictly geometric relationship, calculated by first determining the fraction of the earth over which any one instrument can detect an event (or signals from the satellite can be received). The fraction of the earth area viewed by a satellite deployed instrument, based on the surface area of a unit dome is: Fraction of Earth Area in View =
(1 − cosβ) 2
(1)
The Law of Sines is used to derive β, which is given by: β = cos−1
re cosθ − θ rs
(2)
3 Probabilistic Coverage For GPS-based capabilities, the ratio of earth and satellite orbital radii, r e /r s is equal to 6378 km/(20,200 km + 6378 km) = 0.24. For N total satellites in a constellation, each with a detonation detection instrument or navigation transmitter onboard, assumed to be randomly and uniformly distributed around the sphere of the earth, the average number of satellites in view can be derived as the fraction of the earth area covered by each sensor, multiplied by N: Average Satellites in View = N
1 − cosβ 2
(3)
After inserting β into the above equation and replacing θ with 90-MLA, one obtains a direct relationship between average number of satellites in view over the whole earth and Maximum Look Angle: N −1 r e 1 − sin cos Average Satellites in View = (MLA) + MLA 2 rs
(4)
Our analysis uses the result of Eq. (4) rounded to the nearest integer as an estimate of the actual number of satellites in view so that the rest of the calculations can be done using a hypergeometric distribution. We will subsequently show by comparisons
138
V. H. Standley et al.
with more precise calculations (discussed below) that this assumption is reasonably accurate for the parameters of the GPS constellation. A simple random model can be based solely on mixing the average number of satellites in view from Eq. (4) with those not in view. Confining ourselves first to the half-earth case, we use the hypergeometric distribution with population size N = 12, sample size n = 12, the number of successes in the sample = k, and number of successes in the population = K = Eq. (4). For a specific example where MLA = 90°, K = 9. The mean of the hypergeometric distribution is n * K/N, which is for this case 12 × 9/12 = 9. While not particularly useful, the half-earth case is instructive because it shows that the hypergeometric distribution returns a probability of one for k = K and zero otherwise. That is, the number of satellites in view as calculated by Eq. (4) is always true. A whole earth, 24-satellite GPS constellation can be thought of as twice a half-earth constellation with N = 2 * 12 = 24 and number of successes twice the number for the half-earth case. However, we keep the sample size n equal to 12 and thus the hypergeometric will return results other than simply zero or one. We use 12 because it corresponds to the maximum number of satellites that can ever be in view. Conversely, it can be thought of as how many GPS-based detonation detection instruments are in view of an event on the ground if the instruments had the same look angle for the event. Consider the example of a GPS hand-held receiver that has the equivalent of a satellite sensor Maximum Look Angle of 85°. The integer number of satellites in view for half-earth according to Eq. (4) is equal to 8. For the whole earth, there is an average of 2 * 8 = 16 in view. For reasons explained, we keep the sample size n equal to 12. The probability that exactly eight satellites will be in view using integer values derived from Eq. (4) is: P(X = 8) =
K k
N−K n−k N n
=
16 8
24 − 16 12 − 8 24 12
=
12, 870 × 70 = 0.33 2, 704, 156 (5)
a where N = 24, K = 16, n = 12, k = 8 and is the binomial coefficient for b integers a and b. The result above makes sense because the ratio of the average satellites in view to the total number of satellites, 8/24, is numerically the same. Cumulative probability is more useful because, for example, hand-held receivers need information from at least four satellites to compute a three-dimensional coordinate. The cumulative probability associated with a minimum number of satellites in view is given by:
Earth Coverage Model for GPS-Like Capabilities
139
P(X ≥ k) = 1 −
k−1
K k
0
N−K n−k N n
(6)
Using this model, the probability that at least four satellites are in view of a GPS hand-held receiver for the example above is: ⎛ ⎜ ⎜ P(X ≥ 4) = 1 − ⎜ ⎜ ⎝
16 0
24 − 16 12 − 0 24 12
+
16 1
24 − 16 12 − 1 24 12
+
16 2
24 − 16 12 − 2 24 12
+
16 3
= 1 − (0 + 0 + 0 + 0) = 1
24 − 16 12 − 3 24 12
⎞ ⎟ ⎟ ⎟ ⎟ ⎠
(7)
The average log per capita energy consumption for win, lose, initiator, and noninitiator categories are summarized in Table 1. The weighted average of the log per capita energy for the win cases is −0.42. The weighted average of the log per capita energy for the lose cases is −0.82. On average, those victorious in wars were consuming about 0.4 more in log per capita energy or about 2.5 times more than nations losing wars. Energy consumption is thus an important factor in winning wars whether for attack or defense. This result illustrates the limitation of using a discrete random model. Under some circumstances, the results will either be exactly one or zero, which might leave some doubt about how different the true value will be. However, this concern can easily be addressed by calculating the result for slightly different coverage circumstances. For example, the probability that at least five satellites are in view is given by: 1820 × 1 = 0.999 P(X ≥ 5) = 1 − 0 + 0 + 0 + 0 + 2, 704, 156
(8)
Table 1 Ground station latitudes represented latitudes and Earth surface fractions, and cumulative Earth surface fraction Ground station latitude (°)
Latitude range represented (°)
Fraction of earth surface represented
Cumulative fraction from 0 to 90° latitude
90
80–90
0.02
1.00
70
65–80
0.08
0.98
60
50–65
0.14
0.90
40
36.5–50
0.17
0.76
30
27.75–36.5
0.13
0.59
22.5
15–27.75
0.21
0.46
7.5
3.75–15
0.19
0.25
0.0 (Equator)
0–3.37
0.06
0.06
140
V. H. Standley et al.
That is, there is a 99.9% chance of finding one more than the minimum number of GPS satellites needed to compute a location from anywhere on the earth. Thus, the true value for at least four satellites in view is somewhere between 0.999 and 1, which should be satisfactory knowledge under most circumstances.
4 Validation A method of modeling the number of GPS satellites with sensors in view of an event was also performed by a Monte Carlo sampling of GPS orbital almanac data [2]. This method directly calculates the average number of satellites in view and was used to independently validate the simple hypergeometric analytical model developed. To this end, a dynamic model of the GPS constellation orbits has been represented by randomly sampling an almanac of satellite elevations over time using ephemeris data for a given GPS constellation. GPS satellites exhibit an orbital period of one-half a sidereal day, i.e., 11 h and 58 min. The orbits are arranged so that at least six satellites are always within line of sight from almost everywhere on earth’s surface. Orbiting at an altitude of approximately 20,200 km; orbital radius of approximately 26,600 km, each SV (space vehicle) makes two complete orbits each sidereal day, repeating the same ground track each day. Trimble Planning software, version 2.90, was used to generate lookup tables for the elevation angles of satellites in a constellation of 24 GPS satellites every minute over a 72-h period for a set of assigned “ground station” latitudes. Since the GPS satellite ground tracks repeat each day, use of a three-day lookup table as a repeating pattern over time is deemed a reasonable approximation of true constellation behavior. Further, it is assumed that the southern hemisphere observations of satellite elevations will be a mirror image of those in the northern hemisphere. The latitudes chosen were 0.0, 7.5, 22.5, 30, 40, 60, 70 and 90°, all at 77° longitude. Associated with each value of latitude is a section of the hemisphere of the earth’s surface, each of which covers a fraction of the earth’s surface area, respectively, as shown in Table 1. An almanac of satellite elevations, output at one-minute intervals over a 72-h period, as seen from each of these latitudes is then inputted as a separate ‘sheet’ in one Excel spreadsheet book. A random number generator is used to select one of the 4321-time slots that an event is assigned to take place. The table entries contain all the visible satellite elevations above the horizon at that location. It is then a simple matter to subtract these elevations from 90° to convert to equivalent satellite look angle represented at that location. Next, the spreadsheet randomly assigns a satellite number, 1–24 to the table entries. The converted look angles associated with satellites are then compared to the assumed sensor Maximum Look Angle to determine whether the satellite sensor “sees the event” or not. Of course, the converse is represented as well—whether a ground-based sensor successfully receives a signal from the satellite. If the event is seen or the signal is received on the ground, the detection probability, here assumed to be one, is then multiplied by the input satellite/sensor reliability, say 0.98. Thus, each satellite deemed to have seen an event or had its
Earth Coverage Model for GPS-Like Capabilities
141
timing signal successfully received has an individual detection probability of 0.98. Next, the cumulative probabilities are calculated for the cases of one, two and so on, up to 14 satellites detecting an event. The cumulative probability is defined as Cumulative Probability of Detection = Pdn + Pdn+1 − Pdn × Pdn+1
(9)
(where Pdn is the initial Probability of Detection and Pdn+1 is the additional Probability of Detection from the next satellite) and is calculated recursively to include all satellites detecting an event in the final value of Probability of Detection.
5 Results The relationship between sensor Maximum Look Angle and average number of satellites in view can be calculated for each assumed sensor Maximum Look Angle using a pure geometric calculation and the Monte Carlo sampled almanac results for the GPS constellation. The results are shown in Fig. 2. The assumption of a uniform distribution of satellites around the earth as a representation of the GPS constellation is, in general, a good one, but is better for higher values of sensor maximum look angles. Note, for instance, that the average number of satellites in view suffers at the lower maximum sensor look angles starting at 45° and below. The global average values are good representations for all maximum look angles, but the higher latitudes
Average Number of Satllites in View
10
8
6
4
2
0
0
10
20
30
40
50
60
70
80
Sensor Look Angle
Fig. 2 Satellites in view versus Maximum Look Angle for a 24-satellite constellation
90
142
V. H. Standley et al.
begin to show degradation in performance for almanac-based satellites in view as compared to the pure geometric model. This is caused by the fact that at the North Pole, the actual GPS constellation never rises above about a 45° elevation angle—in other words, the assumption of uniform satellite distribution in the simple geometric model breaks down over the poles for lower values of sensor maximum look angles. The approximation of uniform coverage is reasonable when larger sensor maximum look angles are used because there is then significant overlap between actual satellite orbits overlooking the earth’s surface. From this analysis, it is believed that only for a case of considering a specific place on earth, or for very small sensor maximum look angles, would one need to consider applying a correction factor for latitude. Comparing the probability density functions (PDFs) and cumulative distribution functions (CDFs) in terms of satellites in view provides a clear demonstration that coverage provided by GPS-based capabilities is approximately random. Figure 3 compares the PDFs between the random model and almanac data for an instrument having a look angle between ten and 90° and demonstrates that the two methods show good agreement. DIORAMA (Distributed Infrastructure Offering Real-time Access to Modeling and Analysis) is a software resource developed by Los Alamos and Sandia National Laboratories for the NNSA that enables, in part, estimation of the performance of space-based sensors, including those hosted on GPS satellites. It uses detailed ephemeris associated with the GPS constellation and incorporates a simple Maximum Look Angle model for events near the surface of the earth. Between 50 and 90°, results are all within about 10% of each other. Thus, the simple random distribution approximation using a sample size equal to half the total constellation size should be enough for estimating coverage associated with GPS-based navigation satellites. See Fig. 4.
Fig. 3 Probability density curves from the hypergeometric distribution, Almanac data, and the DIORAMA
RMS Difference
Earth Coverage Model for GPS-Like Capabilities
143
0.25
RMS HG-Almanac RMS HG-DIORAMA RMS DIORAMA-Almanac
0.15 0.05 -0.05 0
10
20
30
40
50
60
70
80
90
Look Angle (Degrees) Fig. 4 Root-mean square (RMS) differences between Hypergeometric (HG), Almanac, and DIORAMA
6 Risk Analysis Application As the number of satellites in the constellation increases, the law of large numbers means that the random approximation is more accurate. A low earth orbit (LEO) navigation capability using GPS-like technology would require many more satellites. Just like GPS satellites, LEO navigation satellites would transmit precise timing to near-surface equipment running a Kalman filter [3] where a minimum of four signals is needed to compute location. Figure 5 shows the probability of receiving less than four signals from satellite’s overhead as a function of Maximum Look Angle, which represents the risk of a ground receiver failing to calculate a position solution. An LEO-based navigation capability would enjoy the benefit of lower power requirements due to shorter receiver-satellite distances, better coverage amidst building or mountains because of more oblique satellites, and greater resilience to anti-satellite threats [4] for sheer numbers. Finally, because costs are dominated by the number of fielded spacecraft, the method may also be used to estimate the cost and cost risk [5]. The method is not limited to the case of space-based navigation where four or more satellites must be in view. The model can be applied, for example, to any capability 1
Log10 P(Satellites