Intelligent Agriculture: Developing a System for Monitoring and Controlling Production (Emerald Points) 9781789738469, 9781789738438, 9781789738452, 1789738466

Intelligent Agriculture: Developing a System for Monitoring and Controlling Production examines the development of Wirel

161 71 8MB

English Pages 201 [208]

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Front Cover
Intelligent Agriculture: Developing a System for Monitoring and Controlling Production
Copyright Page
Dedication
Contents
List of Figures
List of Tables
About the Authors
Introduction
Acknowledgements
Chapter 1 Consumer Discovery
1.1. Introduction
1.2. Benchmark Analysis Against Other Similar Technologies
1.3. Technological Roadmap
1.3.1. Evolution of Cultivation Methods
1.3.2. New Technologies
1.3.3. Advantages and Disadvantages in Technological Advances in Agriculture
1.4. Intellectual Property Analysis
1.4.1. What Intellectual Property Rights Are Important to Protect Computer Programs?
1.4.1.1. The Benefits of Patent Protection
1.4.1.2. Response to Criticisms of Software Patents
1.5. Value Proposal
1.5.1. Broccoli Producers in Mexico
Chapter 2 System Architecture
2.1. Introduction
2.2. Wireless Sensor Network
2.2.1. Sensors
2.2.1.1. Solar Radiation Sensor SP-212
2.2.1.2. Humidity and Temperature Probe Sensor HMP60
2.2.1.3. Soil Temperature and Moisture Sensor 5TM
2.2.1.4. Electrical Conductivity Sensor GS3
2.2.1.5. Soil Moisture Sensor EC-05
2.2.2. Sink Nodes
2.2.3. Solar Panels
2.3. Hardware Brain
2.3.1. Raspberry Pi
2.3.2. Arduino Mega 2560
2.3.3. XBee PRO Series 1
Chapter 3 Software and Applications
3.1. Introduction
3.2. Software
3.3. Software Design
3.4. Application Screenshots
Chapter 4 Agrotechnology Solutions
4.1. Introduction
4.2. Aim and Objectives
4.3. Proof-of-concept Activities
4.3.1. Measuring Outside Temperature and Humidity
4.3.2. Measuring Electrical Conductivity, Moisture, and Temperature of the Soil
4.3.3. Measuring Global Radiation
4.4. Demonstration of Smart Farming Network
4.4.1. Software Used during Network Testing
4.4.2. Graphical Representation of Results Collected from Each Sensor Used during the Demonstration
4.4.3. Software for Setting Real-time Clock
4.5. Feasibility Analysis including Protocol Updates and Proof to Continue with TRL4 and Beyond
Conclusions
Experimental Setup
Results
References
Index
Recommend Papers

Intelligent Agriculture: Developing a System for Monitoring and Controlling Production (Emerald Points)
 9781789738469, 9781789738438, 9781789738452, 1789738466

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

INTELLIGENT AGRICULTURE

This page intentionally left blank

INTELLIGENT AGRICULTURE Developing a System for Monitoring and Controlling Production BY

GONZALO MALDONADO-GUZMÁN Universidad Autónoma de Aguascalientes, Mexico

JOSE ARTURO GARZA-REYES University of Derby, UK

LIZETH ITZIGUERY SOLANO-ROMO Universidad Autónoma de Aguascalientes, Mexico

United Kingdom  North America  Japan India  Malaysia  China

Emerald Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2019 Copyright r 2019 Emerald Publishing Limited Reprints and permissions service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. Any opinions expressed in the chapters are those of the authors. Whilst Emerald makes every effort to ensure the quality and accuracy of its content, Emerald makes no representation implied or otherwise, as to the chapters’ suitability and application and disclaims any warranties, express or implied, to their use. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-78973-846-9 (Print) ISBN: 978-1-78973-843-8 (Online) ISBN: 978-1-78973-845-2 (Epub)

ISOQAR certified Management System, awarded to Emerald for adherence to Environmental standard ISO 14001:2004. Certificate Number 1985 ISO 14001

We dedicate this book project to our families. Their love and constant and unconditional support have been an invaluable source of strength and inspiration to complete this project. Gonzalo Maldonado-Guzmán Jose Arturo Garza-Reyes Lizeth Itziguery Solano-Romo

This page intentionally left blank

CONTENTS List of Figures

ix

List of Tables

xiii

About the Authors

xv

Introduction

xvii

Acknowledgements

xxiii

1. Consumer Discovery 1.1. Introduction 1.2. Benchmark Analysis Against Other Similar Technologies 1.3. Technological Roadmap 1.3.1. Evolution of Cultivation Methods 1.3.2. New Technologies 1.3.3. Advantages and Disadvantages in Technological Advances in Agriculture 1.4. Intellectual Property Analysis 1.4.1. What Intellectual Property Rights Are Important to Protect Computer Programs? 1.5. Value Proposal 1.5.1. Broccoli Producers in Mexico 2. System Architecture 2.1. Introduction 2.2. Wireless Sensor Network 2.2.1. Sensors

vii

1 1 4 13 13 19 23 28

30 34 46 49 49 52 55

Contents

viii

2.3.

2.2.2. Sink Nodes 2.2.3. Solar Panels Hardware Brain 2.3.1. Raspberry Pi 2.3.2. Arduino Mega 2560 2.3.3. XBee PRO Series 1

65 67 73 74 76 77

3. Software and Applications 3.1. Introduction 3.2. Software 3.3. Software Design 3.4. Application Screenshots

81 81 84 98 101

4. Agrotechnology Solutions 4.1. Introduction 4.2. Aim and Objectives 4.3. Proof-of-concept Activities 4.3.1. Measuring Outside Temperature and Humidity 4.3.2. Measuring Electrical Conductivity, Moisture, and Temperature of the Soil 4.3.3. Measuring Global Radiation 4.4. Demonstration of Smart Farming Network 4.4.1. Software Used during Network Testing 4.4.2. Graphical Representation of Results Collected from Each Sensor Used during the Demonstration 4.4.3. Software for Setting Real-time Clock 4.5. Feasibility Analysis including Protocol Updates and Proof to Continue with TRL4 and Beyond

111 111 116 118 120 127 136 141 143

148 155 157

Conclusions Experimental Setup Results

163 164 166

References

171

Index

179

LIST OF FIGURES

Chapter 2 Figure 1.

Structure of WSN. . . . . . . . . . . . . .

54

Figure 2.

Overall System Architecture of WSN. . . . .

56

Figure 3.

Structure of the Sink Node. . . . . . . . . .

66

Figure 4.

Block Diagram of a Sink Node.. . . . . . .

68

Figure 5.

IEEE 802.15.4 Stack. . . . . . . . . . . .

79

Figure 6.

Business Use Case. . . . . . . . . . . . .

86

Figure 7.

Prediction Crop Use Case. . . . . . . . . .

87

Figure 8.

Display Historic Data Use Case. . . . . . .

87

Figure 9.

Manage Sensor Use Case. . . . . . . . . .

88

Figure 10. Filling Database Use Case.. . . . . . . . .

88

Figure 11. Employee Administration Use Case. . . . . .

89

Figure 12. Prediction Menu Sequence.. . . . . . . . .

89

Figure 13. Crop Menu Sequence. . . . . . . . . . . .

90

Figure 14. Manage Sensor Sequence.. . . . . . . . .

91

Figure 15. Filling Database Sequence.. . . . . . . . .

91

Figure 16. Employee Management Sequence. . . . . .

92

Chapter 3

Figure 17. Entity-Relationship (ER) Database. . . . . . . 100

ix

x

List of Figures

Figure 18. Administrator/User Access.. . . . . . . . . 101 Figure 19. Administrator Menu. . . . . . . . . . . . . 102 Figure 20. Predictive Menu. . . . . . . . . . . . . . . 102 Figure 21. Formula Menu Button. . . . . . . . . . . . 103 Figure 22. Maximum Rainfall Bar Graph. . . . . . . . 103 Figure 23. Rainfall Probability Dispersion Graph for the Next Seven Days. . . . . . . . . . . . . . 104 Figure 24. Rainfall Probability Pie Graph for the Next Seven Days.. . . . . . . . . . . . . . . . 104 Figure 25. Cosecha (Crop) Menu Option. . . . . . . . 105 Figure 26. Sensor Menu Option. . . . . . . . . . . . 105 Figure 27. Employee Button Menu. . . . . . . . . . . 106 Figure 28. “Agregar Personal Nuevo” Button. . . . . . 106 Figure 29. “Editar Personal” Button. . . . . . . . . . . 107 Figure 30. “Eliminar Personal” Button. . . . . . . . . . 107 Figure 31. “Ayuda” Button (Help). . . . . . . . . . . . 108 Figure 32. Normal User Menu. . . . . . . . . . . . . 109 Chapter 4 Figure 33. Smart Farming Zigbee Cluster Tree Network.

118

Figure 34. Example of Comma-separated Value. . . . . 119 Figure 35. Flowchart for HMP-60 Sensor Communication with Router. . . . . . . . . . . . . . . . . 122 Figure 36. Pin Assignments, Configuration Values, and Configuration Bits of HMP-60 Sensor. . . . . 124 Figure 37. Main Code of HMP-60 Sensor.. . . . . . . 125 Figure 38. Raspberry Pi Software. . . . . . . . . . . . 126 Figure 39. Raspberry Pi, HMP-60 Received. . . . . . . 127

List of Figures

xi

Figure 40. Graphical Representation of HMP-60 Testing. 128 Figure 41. Flowchart of Software for 5TE Sensor Probe. . 130 Figure 42. Header Files, Configuration Values, and Configuration Bits of 5TE Sensor. . . . . . . 131 Figure 43. Main Code for 5TE Soil Sensor (A). . . . . . 132 Figure 44. Main Code for 5TE Soil Sensor (B). . . . . . 133 Figure 45. Sensor Measurement of Dry Plant. . . . . . . 134 Figure 46. Sensor Measurement of Watered Plant. . . . 135 Figure 47. Raspberry Pi, 5TE Sensor Data Received. . . 136 Figure 48. Graphical Representation of 5TE Sensor Data Received. . . . . . . . . . . . . . . . . . 136 Figure 49. Flowchart of SPI-212 Global Radiation Sensor Arduino Software. . . . . . . . . . 138 Figure 50. Configuration Bits of SPI-212 Sensor Software on Arduino. . . . . . . . . . . . . . . . . 139 Figure 51. Main Code of SPI-212 Sensor Software on Arduino. . . . . . . . . . . . . . . . . . 139 Figure 52. Raspberry PI, SPI-212 Sensor Data Received.

140

Figure 53. Graphic Representation of SPI-212 Sensor Data Received. . . . . . . . . . . . . . . 141 Figure 54. Small Scale of Network. . . . . . . . . . . 142 Figure 55. Example of Simple Bridge Connection. . . . 142 Figure 56. Software on the Coordinator. . . . . . . . . 144 Figure 57. Software on the Bridges. . . . . . . . . . . 145 Figure 58. Data Send by 5TE Soil Sensor. . . . . . . . 147 Figure 59. Data Send from SPI-212 Global Radiation Sensor. . . . . . . . . . . . . . . . . . . 148 Figure 60. Data Send from HMP-60 Outdoor Temperature and Humidity Sensor. . . . . . 149

xii

List of Figures

Figure 61. Data Collected on Coordinator. . . . . . . 150 Figure 62. Graph of Outdoor Temperature vs Time.. . . 151 Figure 63. Graph Representing Outdoor Humidity vs Time. . . . . . . . . . . . . . . . . . . . 152 Figure 64. Graph Representing Global Radiation. . . . 153 Figure 65. Graph Representing Soil Dielectric Constant vs Time. . . . . . . . . . . . . . . . . . . 154 Figure 66. Graph Representing Soil Moisture vs Time.. . 154 Figure 67. Graph Representing Soil Temperature vs Time. . . . . . . . . . . . . . . . . . . . 155 Figure 68. Graph Representing Temperature and Humidity vs Time. . . . . . . . . . . . . . 156 Figure 69. Software for Setting Time and Data on Real-time Clock. . . . . . . . . . . . . . . 157 Figure 70. Software for DTH-22 Temperature and Humidity Sensor.. . . . . . . . . . . . . . 158 Figure 71. Technology Readiness Levels. . . . . . . . . 159 Conclusions Figure 72. Smart Farming Wireless Sensor Network. . . 165 Figure 73. Smart Farming Wireless Sensor Network Communication. . . . . . . . . . . . . . . 166 Figure 74. Arduino Mega 2560 Serial Monitor-A. . . . 167 Figure 75. Graphical Representation in Raspberry Pi (Python Shell). . . . . . . . . . . . . . . . 168 Figure 76. Sensor Reading from the Sensor Node Plotted Graphically.. . . . . . . . . . . . . . . . 168

LIST OF TABLES

Chapter 1 Table 1.

Companies that Offer Agrotechnology Services. . . . . . . . . . . . . . . . . .

14

Table 2.

The Most Important Sowings in Mexico. . . .

43

Table 3.

Broccoli Producers in Mexico. . . . . . . .

44

Technical Specification of Atmospheric Sensors.. . . . . . . . . . . . . . . . . .

63

Table 5.

Technical Specification of Soil Sensors. . . .

64

Table 6.

Battery Power Consumption Calculation. . .

71

Table 7.

Arduino Runtime from V44 Batteries. . . . .

72

Table 8.

Technical Specification of Solar Panel and V44 USB Battery. . . . . . . . . . . . . .

Chapter 2 Table 4.

Hardware Components. . . . . . . . . . .

72 73

Table 10. Raspberry Pi3 Model B Technical Specification. . . . . . . . . . . . . . . .

75

Table 11. Arduino Mega 2560 Technical Specification.

78

Table 9.

Chapter 3 Table 12. Actor Description/Functions. . . . . . . . .

85

Table 13. Prediction User Case Description. . . . . . .

93

xiii

xiv

List of Tables

Table 14. Crop User Case Description. . . . . . . . .

94

Table 15. Manage Sensor User Case Description. . . .

95

Table 16. Filling DB User Case Description. . . . . . .

96

Table 17. Employee Administration User Case Description. . . . . . . . . . . . . . . . .

97

Chapter 4 Table 18. Smart Farming Measurement Frequency.. . .

119

Table 19. Connections between HMP-60 Sensor and Arduino Mega 2560. . . . . . . . . . . .

121

Table 20. Description of Devices Used. . . . . . . . .

146

ABOUT THE AUTHORS Gonzalo Maldonado-Guzmán is a Professor at the Universidad Autónoma de Aguascalientes, Director of the Small and Medium Enterprises Observatory, and Director of the Research and Postgraduate Studies Department. His areas of research include marketing, corporative social responsibility, innovation and knowledge management, and IT and intellectual property in small and medium size enterprises (SMEs). He has coordinated projects in the Aguascalientes state, Mexico, in innovation and organizational culture in micro and SMEs. He has international projects with Universities of Murcia, Cantabria and Cartagena, in Spain. Jose Arturo Garza-Reyes is a Professor of Operations Management and Head of the Centre for Supply Chain Improvement at the University of Derby, United Kingdom. He is actively involved in industrial projects where he combines his knowledge, expertise, and industrial experience in operations management to help organizations achieve excellence in their internal functions and xv

xvi

About the Authors

supply chains. He has also led and managed international research projects funded by the European Union, British Academy, British Council, and Mexico’s National Council of Science and Technology (CONACYT). As a leading academic, he has published over 100 articles in leading scientific journals, participated in international conferences, and has four books in the areas of operations management and innovation, manufacturing performance measurement, and quality management systems. Professor Garza-Reyes is Associate Editor of the International Journal of Operations and Production Management and Journal of Manufacturing Technology Management as well as the Editor of the International Journal of Supply Chain and Operations Resilience and Editor-in-Chief of the International Journal of Industrial Engineering and Operations Management. The areas of expertise and interest for Professor Garza-Reyes include general aspects of operations and manufacturing management, business excellence, quality improvement, and performance measurement. Lizeth Itziguery Solano-Romo is a Professor at the Universidad Autónoma de Aguascalientes. Her areas of research include information technology management, IT use and adoption, and digital marketing in SMEs. She has participated in the Aguascalientes state, Mexico, in the implementation of the new criminal justice system. She has international project participation to reduce the IT gap between public and private universities (ALFA-EU) with Universities of Finland, Romania, Brazil, Ecuador, and Colombia.

INTRODUCTION Agriculture is today one of the fields of knowledge least analyzed and discussed by various researchers, academics, and professionals not only in the field of agriculture but also in different areas of knowledge, although it is an elementary construct for the existence of humanity itself (Ding et al., 2018). Also, currently, the total world population amounts to a little more than seven billion people, and according to the estimates that have been made by the main international organizations, it is expected that by the year 2050, it will generate a substantial population growth of a little more than 2.5 billion people, which will be located primarily in the main urban cities, which will mean that a little more than 90% of the total world population will be concentrated practically in two continents: Asia and Africa (Lloyd, 2017). However, world food production is totally limited, especially in Africa, and the serious problem of food shortages worldwide has not yet been resolved (Sánchez, 2002). In addition, the Asian continent has serious problems of shortage of drinking water (Pomeranz, 2009), even though 72% of the total surface of the earth is covered by water, and it is estimated that there are a little more than 1.45 billion cubic kilometers of water. Despite the existence of an extensive territorial extension covered by water, a little less than 1% of the total water on the planet is fresh water that is used not only for human consumption but also for agricultural xvii

xviii

Introduction

irrigation, which represents a little more than 13 billion hectares; however, only 22% of that land is potentially arable (Lal, 1990). In this context, there are currently diverse countries that apply traditional agriculture methods that have a high consumption of potable water, are intensive in labor, use fungicides and pesticides that are highly polluting, and are low in productive efficiency (Ding et al., 2018). Therefore, considering the significant increase in the world’s population, the severe shortage of drinking water, the existing limitation of resources, and the low level of efficiency of agricultural productivity, among other factors, it is indispensable and urgent that researchers, academics, and professionals from all areas of scientific knowledge guide their studies in the analysis and discussion, not only of the efficiency of a regulated agriculture but also in the development of agrotechnology that propitiates an Intelligent Agriculture, because this will allow an adequate utilization of the available resources. In this sense, even when the systems of Smart Agriculture are too complex, multivariate, and unpredictable (Kamilaris, 2018), it is also possible to incorporate classic technological controls, such as integral processes or differentiated integral processes (Christofides, 2013; Afram and Janabi-Sharifi, 2014), which are not only easy to implement but also to control the movement processes they generate, thereby allowing an adjustment in the control of energy and the time of consumption (Wang, 2001). In addition, the use of intelligent methods such as the control of fuzzy logic, linear regression, and artificial neural networks involves not only deterministic mathematical models but also generalized mathematical models and mixed models, which allow the development of predictive models of agricultural production more accurately (Afram and Janabi-Sharifi, 2014).

Introduction

xix

Likewise, the use of these mathematical methods require a high level of reasoning and understanding and are generally based on the use of historical data on agricultural or agroindustrial production, or on the generation of expert or high-level knowledge (Ding et al., 2018). Therefore, the performance of the mathematical models of control and prediction of agricultural production is superior to that of the classic models of production control, and they are generally simpler to implement when using intelligent algorithms through computers. Thus, the mathematical models of production control and prediction have a high reliability and accuracy of the levels of agricultural and agroindustrial production, in addition to significantly reducing the use of drinking water, electricity, and emission of CO2 (Ding et al., 2018). Similarly, control and prediction models of agricultural or agroindustrial production generally refer to the use of advanced algorithms through computers that are used to explain and develop predictive models of future growth that plants will have, or the growth that is estimated to have food production (Qin and Badgwell, 2003). Therefore, this type of control and prediction models work with a series of inputs that are controlled by the computers during a certain period of time, and they take the data usually from a selected sample of a dataset that reveals agricultural or agroindustrial production; however, only some of these models are implemented in the production prediction process (Bumroongsri and Kheawhom, 2014) because they generate the smallest possible error in the prediction of food production. In addition, the use of advanced algorithms in the models of control and prediction of agricultural production is often done through three steps: prediction models, optimization in its implementation, and adjustment in the feedback (Zhang, 2017), with these three steps being equally important for the development of agricultural control and prediction models.

xx

Introduction

Production control and prediction models were developed at the beginning of the 1960s, and these types of models were used almost exclusively in the process of predicting industrial production (Garriga and Soroush, 2010); however, its use has expanded to all areas of scientific knowledge, and its use has been considered important and paramount in all production prediction processes, including, of course, agricultural and agroindustrial production. Additionally, most of the production control and prediction models require a series of constraints, predictive information, and linear and nonlinear dynamics for their application (Ding et al., 2018). Linear models of control and production prediction are usually used to solve quadratic problems of online programming, and nonlinear production control and prediction models are generally used to control systems with nonlinear dynamics, for which undoubtedly greater mathematical calculations than linear models (Vukov, 2015) are required. In addition, matrices of control dynamics and controlled algorithm models, which are commonly based on linear quadratic mathematical models that are relatively easy to use, have recently been incorporated into the theory of production control and prediction models. Within the models of controlled algorithms are the models of internal control, which are widely used by researchers, academics, and professionals in the field of computer science and mathematics, and which can be defined as a simple entry and/ or exit of information through a discrete time series system (García and Morari, 1982). Therefore, it is possible to affirm that the internal control models are nothing more than a combination of a dynamic control matrix and a model of control algorithms, but theoretically it is better; and the internal control model is more complete than the two previous models, and usually the internal control model tends to solve the problems of control and production prediction more

Introduction

xxi

robustly and with a much smaller error, which makes the model more efficient and effective. Therefore, given that industrial processes are increasingly complex, involve an increasing number of interfaces, and are strongly non-linear, it is essential that new production control and prediction models are adapted and implemented in the companies of all sizes and sectors, as is the case of internal control models, which are more robust and have the minimum possible error in their application (Ding et al., 2018). However, the time to perform the calculations for the internal control models should be relatively long and totally efficient, to aspire to obtain robust results and with a minimum error, for which researchers and academics have considered necessary that this type of models be stabilized (Ding et al., 2018), that is, that they adapt to the production processes of the companies where they will be applied (Zhang, 2017).

This page intentionally left blank

ACKNOWLEDGEMENTS We thank the British Council for having financially funded the international research project entitled Developing Food Security and Water Conservation for Economic Growth in Mexico  A Smart Monitoring and Control System (SMCS) Agro-Technology for Sustainable and Efficient Farming Operations (No. 275317449), from which this work is derived. The project was funded through the Newton Fund and the Institutional Links scheme of the British Council, and it was carried out through an international collaboration between the University of Derby (UK) and the Universidad Autónoma de Aguascalientes (Mexico). We would like to thank our institutions, the University of Derby (UK) and the Universidad Autónoma de Aguascalientes (Mexico), for their unconditional support to complete the research project and production of this book. Also, we would like to thank our publisher “Emerald Publishing Limited” and its editorial team for assisting us with this publication. Finally, we would like to express our deepest gratitude to our following colleagues who also made a significant contribution to the research project and this work: • Dr Jose Manuel Andrade, Senior Lecturer in Electrical And Electronic Engineering, University of Derby, UK.

xxiii

xxiv

Acknowledgements

• Gisha Gangadharan, Research Assistant Engineer in Electrical and Electronic Engineering University of Derby, UK. • Christopher Horry, Student Research Assistant in Electrical and Electronic Engineering, University of Derby, UK. • Ruben Michael Rodríguez-González, Student Research Assistant in MBA, Universidad Autónoma de Aguascalientes, Mexico.

CHAPTER 1 CONSUMER DISCOVERY

1.1. INTRODUCTION The literature pertaining to business studies and management shows that various researchers, academics, and professionals belonging to that field are paying special attention to the use of data and information that reflects the changes that are being made to the current or future assets or services of the consumers and suppliers (Chien-Hsing et al., 2005). Therefore, it is extremely important for today’s organizations to process all information regarding the use of technology by consumers. This would facilitate the use of currently available information and technology, and increase the value of future applications that would be developed. This, in turn, could generate a higher level of acceptance among current consumers, since new applications would not only facilitate the use of current technology but also increase its usefulness. Consequently, the extensive development of technological support in decision-making processes that are based on data and information has increased significantly in the last two decades (Chien-Hsing et al., 2005). One of the pioneering works in this regard has been conducted by 1

2

Intelligent Agriculture

Shim et al. (2002), who presented a descriptive analysis, through an extensive review of the literature, of the technological support prevailing in the decision-making capacity of the companies, with regard to their past, present, and future implementation. Thus, to use the existing data and information more efficiently with respect to potential consumers, new information technologies have introduced new developments and applications, technologically developed products, and new knowledge generated through the discovery of potential customers, which allows the creation of databases and its data mining in organizations (Chien-Hsing et al., 2005). In this sense, databases and data mining fundamentally orient the integration of information management of current and future clients, are a representation of knowledge, and improve the understanding regarding the use of new machinery, equipment, or developed technology (Fayyad and Stolorz, 1997). This generally facilitates the descriptive and predictive processes for the discovery of new clients or business partners through a series of historical data that were previously collected. Therefore, databases and data mining are the two fundamental elements for the discovery of customers in any current business, and its use has been significantly extended in the last two decades by researchers, academics, and professionals of the business sciences as an effective measure for the making and management of decisions (Han and Fu, 1999). Thus, for example, the discovery of new customers who prefer to have a low annual reward will reveal consumers who are more predisposed to buy the product or service that is being offered than those consumers who prefer high levels of rewards; this will facilitate in decision-making (ChienHsing et al., 2005). Pitta (1998) came to the conclusion that the generation of databases and data mining are important

Consumer Discovery

3

tools that marketing personnel of companies can use to discover customers who will be willing to consume their products: products, services, or technology. Feelders, Daniels, and Holsheimer (2000) in their study presented a fundamental concept for the generation of data mining, as well as the processes that are to be followed for their application; this would ensure that the application is carried out properly in a timely manner. In this context, the application and development of databases and data mining in various business areas are increasing not only because of its usefulness but also because of their importance in making business decisions and for the discovery of customers for new products, services, or technology (Chien-Hsing et al., 2005). Thus, the studies oriented regarding the market data of hotels (Sung and Sang, 1998), the prediction of bankruptcy of personnel (Donato et al., 1999), the services of support for the clients, those that are dealt with in the special edition published by Kohavi and Provost (2001) in the Journal of Data Mining and Knowledge Engineering, and budget allocations for the acquisition of library materials (Wu, 2003) are clear examples of studies conducted that consider the generation and use of databases and data mining as essential in the discovery of business customers. In addition, the development and implementation of databases and data mining usually follows six essential steps: (1) identification of the problem, (2) collection of data or information, (3) reprocessing of data or information, (4) processing of data or information, (5) implementation and interpretation of the data or information found, and (6) evaluation of the data or information found. Thus, the identification of the problem helps to specify the criteria or questions that should be commonly raised for the improvement of the decisionmaking capacities. Data collection involves the compilation

4

Intelligent Agriculture

of data and important information that is generally used for the definition of objectives and goals of the organization. The reprocessing of data uses diverse mathematical operations that redefine and reconstruct the consistency of the information, the attributes of the same, and a combination of data and information that are highly correlated with each other in order to reduce the maximum possible errors while decision-making. Data processing generally employs a data mining mechanism to capture the knowledge discovered through association, classification, regression, clustering, and summarization  generalization. The interpretation of discovered or generated knowledge results in its use through different techniques that provide output in the form of text, tables, figures, graphs, animations, diagrams, etc. Implementation is nothing but the transformation of the results obtained while making business decisions. The results generated by the implementation of the knowledge are evaluated through various performance tests. Finally, following the aforementioned six stages for the development and implementation of the databases and according to the requirements of the British Council and the Newton Fund, the discovery of the clients for the development of agrotechnology can usually be done by applying four fundamental elements: (1) benchmark analysis against other similar technologies, (2) technological roadmap, (3) intellectual property analysis, and (4) value proposal. All of these will be analyzed in detail in the following sections.

1.2. BENCHMARK ANALYSIS AGAINST OTHER SIMILAR TECHNOLOGIES In recent years, in Mexico agriculture is gaining importance. In 2007, 20,184 hectares were under plantation. The main

Consumer Discovery

5

states involved in agriculture, in order of importance, are Guanajuato (with more than 50% of its area being planted), Puebla, and Michoacán. In these states, the usage of balers and freezers is very common. Such practices allow the produce to be shipped to other countries in America, Europe, and Asia. Broccoli (Brassica oleracea) is a crop that develops mainly during the autumn and winter seasons. For the plant to develop normally, temperatures during the growth phase must oscillate between 20 and 24ºC. To start the floral induction phase, temperatures between 10ºC and 15ºC are required for several hours of the day. The plant prefers soils that are acidic and not too alkaline, with an optimum level of pH (acidity) between 6.5 and 7. Additionally, the plant also requires soil of a medium texture, which supports an excessive salinity of the soil and a better irrigation of water. Currently, there are varieties that tolerate humid warm climates: Tamer, Avenger, and Maximum, which can adapt to the climatic conditions of southern Sinaloa, especially during the months of November to February (winter). For this reason, it is important to validate the broccoli varieties and determine whether it is feasible to grow them in that particular area as they are widely in demand by the freezing companies; moreover, broccoli cultivation could be an alternative for the horticultural producers of the southern regions of the state. Broccoli is a vegetable that develops favorably in cold and fresh climates, tolerating temperatures of up to 2ºC, as long as the inflorescence is not present in the plant; or else, it will be easily damaged by the drop in temperature. Its optimum development temperature is 17ºC. It adapts almost to any type of soil, but, like all vegetables, it prefers not very light, uniform, deep soils with good drainage and with an optimum pH of 6 to 7.5 (although it grows even from 5 to 5.5 pH). It can be planted directly or transplanted, with the

6

Intelligent Agriculture

latter being currently considered to be the best form of establishment. This is because the new varieties require sowing in trays if 100% advantage is to be obtained from the seeds that are procured from the trading houses. Planting densities are varied: ranging from 30,000 to 80,000 plants per hectare; this will depend on the prevailing climatic conditions of the region and the market where the final product is destined, that is, whether for the industry or for sale in the fresh market. In order to obtain these densities, different planting systems are also managed. In order to obtain a population density of 80,000 plants per hectare, it is necessary to make furrows to each meter and plant in a double row for every 22 cm; in a single row, each 33 cm, 30,000 plants per hectare can be obtained. In plantations of high densities, it is necessary to carry out the transplant in a triangular manner, since in this way space is better utilized and air circulation is much better, favoring the reduction of diseases. Currently, broccoli has diverse varieties with different shapes, colors, and sizes; its grains range from fine to coarse, and from faint green to intense green; likewise, its inflorescence can be from very compact to semi-open. These aspects are taken into consideration in the production regions; there are varieties that can be classified into early, intermediate, and late. Over the years, the desirable characteristics have been selected and their property increased for producers so as to produce varieties with high productivity, resistance to pests and diseases, and adaptation to different types of climates, existing not only in the different regions of the geography of Mexico (north, central, and south) but also in every country of Latin America. In Mexico, the main broccoli-producing states are Guanajuato, Michoacán, Puebla, and Jalisco, with 13,337, 2,509, 1,350, and 1,248 hectares, respectively, with yields

Consumer Discovery

7

ranging between 11 and 21 tons per hectare (t/ha). The prices of broccoli in the national market are diverse, but it usually remains constant almost all the year round: during 2009, they ranged between 12 and 17 Mexican pesos, with an average of 13.86 Mexican pesos per kilogram of fruit (Central de Abastos de México). The price in the US markets ranges from 1.16 to 1.32 dollars per pound of fresh fruit. All the aforementioned data show that broccoli cultivation is a profitable production option for Sinaloa; moreover, it is important to locate new regions for production, so that the export of the product is increased resulting in the generation of a large amount of foreign currency by its sale. Currently, there are diverse companies offering technical and information services to agroindustrial companies, the following are the most important worldwide: (1) VegHands ○ Together with several suppliers, VegHands offers the best of the market in terms of harvest systems  from harvesting bands to harvesting and packing solutions for the field. ○

It specializes in lettuce (iceberg), broccoli, cabbage, herbs, and similar products.



It provides harvesting bands.



It has mobile bands, with single or double band (on which one can also work with boxes). Strong but lightweight bands can be easily mounted next to the trailer, and can be moved along during the harvest for optimal logistics. The bands can be mounted from one trailer to the other in just a few minutes.



Fixed bands are mounted on trailer especially for the transport of boxes and baskets (construction of single

Intelligent Agriculture

8

or double band). They are easy to mount next to the trailer during transport from field to field. ○

Fixed chains are mounted on a trailer and are ideal for unloading boxes and baskets placed in the holder. They supply many alternatives, tailored to one’s needs. It is possible to combine the fixed chains with a special system for rotation of baskets, with which one can change filled baskets without stopping or changing the truck basket.

(2) Brioagrow ○ Brioagrow helps farmers improve the information they have obtained regarding the evolution of their crops by monitoring, in real time, the main variables in which they can intervene. Measuring seven fundamental variables, both environmental and edaphic, together with geolocalized weather forecasts, they allow the farmer achieve maximum production, with the best quality, and reduction in water consumption, fertilizers, and energy. Among the variables controlled by the farmer are soil moisture, at different depths, conductivity (to know the mineral salts of the soil), and temperature; and they help control temperature, humidity, luminosity, and leaf wetness in the environment. They integrate the geolocalized meteorological data plus rain forecasts, cloudiness, relative humidity, and wind speed and direction into their dashboards . ○

The application of Big Data and data mining is fundamental for the detection of diseases in the field and the efficiency to which one can achieve today, as well as in the future, with this new technology.



The more data one has, the greater the degree of reliability on the patterns and resolutions obtained

Consumer Discovery

9

from it. The conclusions drawn are that there are no master formulas. Each farm and each crop has its own peculiarities, and the conclusions drawn from the climatic conditions of an area on the banks of Navarre are hardly similar to those of the middle zone. Therefore, one’s farm, with respect to one’s neighbor’s, has different environmental conditions, which make a disease or pest behave differently. The future is related to the data; the greater the amount of data within the system, the greater the reliability on the predictive algorithms. Moreover, with more sensors, more information, and more historical data, the degree of precision may be greatly improved. (3) HoneyComb ○ Autonomous flight: The AgDrone System™ takes advantage of an advanced autopilot system, allowing the drone to fly itself. No flying experience is required. The user draws a polygon defining the flight area and mission planning software automatically calculates the route. Safety measures come standard including low battery warning and return to home. ○

Cloud data services: Data processing services and tools are necessary to produce actionable information for any agricultural drone. HoneyComb has made the process of capturing and processing data seamless, and is optimized for agricultural drone applications. Advantages include six-channel image processing, one-click data handling, robust algorithms, and optimized hardware configuration.



Optimized for agriculture: HoneyComb has matched its AgDrone System™, MakeMap™ Processing, and

Intelligent Agriculture

10

HoneyComb Farm™ application to produce superior results in agriculture. ○



Multispectral processing: With dual cameras, HoneyComb uses six channels to produce highdefinition maps without sacrificing spectral accuracy or resolution. Automated from start to finish: With HoneyComb, one only has to upload imagery, and it handles the rest. No workflow or geographic information systems (GIS) expertise is required.



Fast processing: As soon as a job hits HoneyComb’s servers, it begins processing. There are no queues or waiting in line.



Visible and normalised difference vegetation index (NDVI) maps: HoneyComb helps generate visible NDVI maps with every flight, eliminating the need to fly twice.

(4) Smart Fertilizer Management Software ○ This software provides accurate and sustainable fertilizer planning. ○

Fertilizers and farm management are managed by data.



It has complete data on nutrient requirements for more than 250 crops.



It easily interprets field test results  analysis of soil, water, and plant tissue.



It recommends fair fertilizer usage.



It has cost optimization modules that improve efficiency and save costs.

Consumer Discovery

11



It helps design nutritive solutions and fertilizer recipes for fertigation and hydroponics.



It balances soil pH by designing liming applications.



It calculates the quality of the water.



It designs fertilizer mixtures and foliar applications.



It uses family units.



It has a built-in warning module.



It has advanced reporting module.

(5) Agrowin ○ Agrowin provides a total management system for agriculture. ○

The agricultural software AgroWin is a system specially designed to help the agricultural entrepreneur in the management, planning, and monitoring of the company and its resources.



Additionally, it allows the reduction of costs and the improvement of income.

(6) Sismagro ○ Satellite mapping: With satellite maps, one can add in the geolocation of one’s lots in less than a minute to determine one’s arable land. In a few clicks and without any necessary training, one can start managing everything that is needed to manage one’s field. ○

Production follow-up: From the map and your crops, the farm management software Sismagro can help to manage agricultural production, keeping traceability of all costs and inputs. It helps to organize one’s calendar to add the work that needs to be done and see in a

Intelligent Agriculture

12

simple way everything that is happening in ’one’s field and lots. (7) Agroptima ○ Agroptima is a revolutionary map that changes the rules of the game. ○

It helps locate one’s fields and crops, as well as consult surfaces and agricultural plots geographical information system (SIGPAC) codes. Thus it is impossible to get lost.

(8) Aguasmarket ○ Aguasmarket is a crop management software. ○

It is a software for water, gas, and electricity management.



It is a powerful management software that ensures better performance of crop management; this irrigation system allows a cross-reference between the parameters to improve the understanding of plant behavior while applying irrigation strategies with good fundamentals.



The system monitors its valves and sensors, indicating exceptions according to its programming pre-adjustment plan.

In the exhaustive search that was carried out in the different databases available on the Internet, it was found that these eight companies offer services and perform activities similar to those that would be had with the development of a new agrotechnology. For this, it is important to establish that some of the services that the agroindustrial companies of Mexico would obtain with the new agrotechnology are already offered in these eight companies worldwide. However, there is no company offering the prediction services of

Consumer Discovery

13

agricultural or agroindustrial production. This would be the differentiating factor between the agrotechnology that is proposed to be developed and the existing services at the global level. Accordingly, the agrotechnology that is to be developed may be developed differently from that already existing in the market and, consequently, some services that are not currently offered by any company worldwide will be added. In this sense, a benchmark was created by comparing those companies that offer similar services worldwide to those that would be developed with the help of the new agrotechnology. Eight companies in which benchmarking may be applied were identified. Each of the eight companies were analyzed in detail in reference to the services they offer to agricultural producers and agroindustrial companies. Table 1 adequately summarizes all the information presented previously and provides a general panorama of the benchmarking that was carried out with other existing companies.

1.3. TECHNOLOGICAL ROADMAP 1.3.1. Evolution of Cultivation Methods The land has been under cultivation for more than 10,000 years. Through time, agricultural practices have changed, but it has its regularities. Regularities are characteristic of any system or technical process. Over time, the prevailing technical system is replaced by a new system that performs the same function. First, the system displays a stage of expansion followed by one of trimming. In the expansion stage, the number of elements, parts, and subsystems of a technical device increases (or, correspondingly, the number of steps of

14

Table 1. Companies that Offer Agrotechnology Services. Company

Cost

VegHands

Not specified

Brioagro

For service

Origin Country Netherland Spain

Technologies HoneyComb Smart Fertilizer

Description

Web Page

Basic tools and training for

https://www.veghands.

collection

com/es

Intensive agriculture, nurseries and

http://brioagro.es

gardens For service For service

Management

USA

Drones, mapping, software and

http://www.

services

honeycombcorp.com

America and

Software for fertilizer and expert

http://www.smartfertilizer.

Europe

advice

com/es

For service

Spain

Software and hardware

www.agrowin.com

Sismagro

Not specified

Latin America

Mapping of lots and fields

www.sismagro.com

Agroptima

Not specified

Spain

Software and applications

https://www.agroptima.com

Aguamarket

Not specified

Latin America

Products for water control

www.aguamarket.com

Software Agropeciario

Intelligent Agriculture

Agrowin

Consumer Discovery

15

an operation process). Then, with evolution, things are combined in an integral construction or set that reduces operations to one, making the device or process simpler. This “expansioncut” pattern can be built for all fields of technology, for example, for the process of planting crops. • Sowing in holes. When people started to cultivate, they did not know about implements. The seeds were thrown into holes made in the ground with the tip of a stick. The disturbance of the ground was minimal. In fact, almost zero. Simple methods were used to control weeds and for fertilizing. Trees were burned to free space for sowing, with the ashes acting as fertilizer. • Plowing without covering the furrow. As the population grew, the need for food was greater. The next step in agriculture was to invent a wooden plow that was pulled by people or animals. Usage of the wooden plow increased the productivity of the crop, but there were almost no changes in the technology pertaining to grain growth itself. Instead of using holes, the seeds were dropped into the furrow that the plow created. The groove was manually closed or by covering the seed with a branch, which is the prototype of the modern harrow. Weed control and fertilizing did not undergo any changes  fortunately, forests were still in abundance. • Plowing. The plow was an effective implement to clean and fertilize new areas while it was possible to burn forests. But this, however, could not be eternal. The fertility of the soils under usage repeatedly degrades. New methods were found to increase land productivity such as loosening soils and controlling weeds. During this time in Ancient Greece, a plow with a mold was used, which plowed under the weeds to depths from which the weeds

16

Intelligent Agriculture

could not emerge again. The classical techniques of those times included plowing with a molded plow, sowing, and manual capping, which although was distressing was much better than felling and burning. • A technique expanded to the maximum to cultivate the soil. An increase in productivity remained the most urgent and the only apparently acceptable demand to combat famine. More and more operations were added to the plow, according to the “expandtrim” trend. The most used expanded technique in the twentieth century included plowing, several growing cycles, leveling, and pre-sowing. The sowing technique is still currently in use. On the other hand, the power of tractors as well as the depth of the plows and their land coverage has shown an steady growth. The same plows were continuously improved. “Soft” plows that left no traces appeared on the scene. Tillage implements were combined in hybrid devices, such as cars, complex and expensive. Huge resources were invested for it, with the usage of oil and labor growing beyond all conceivable limits. As a result, the upper layer of the soil became softened while the deeper layer became compact. In the flat regions, the winds frequently whipped up the fertile layer of soil, leaving behind a desert. The dust storms became the scourge of farmers. Soil erosion extended to tens of millions of farmland. The content of humus in the best lands was reduced from 10%12% to 5%6%. The damage caused by the plow became obvious, with alarms sounding in the most awakened minds. Deep plowing of the soil created great damage. However, one needs to examine: what is harmful with the plow? First, it is a fact that the soil on which the plow works is a

Consumer Discovery

17

complex living organism. Its upper layer is inhabited with aerobic bacteria. Bacteria that live in deep soils are anaerobic and die in the presence of oxygen. When the soil layer is turned, the aerobic bacteria go to the bottom and die due to lack of oxygen, while the anaerobic bacteria are exposed to oxygen and die as well. And it is the life of these bacteria that ensures the fertility of the soil and the accumulation of humus there. Second, the use of the plow results in the formation of a compact soil layer at 2025 cm depth. Under normal conditions, the soil breathes: moisture moves permanently through capillaries. At reduced atmospheric pressure, it is supplied to the uppermost layers of the soil from the deeper layers, while a higher-than-required atmospheric pressure causes the moisture to move to the deepest part of the soil. When the plow destroys the capillarity of the soil, the circulation of moisture stops and the yields go down. • Seeding without plow. It was the American researcher Edward Faulkner who warned the world about the problems related to the plow in his successful book Plowman’s Folly, which recorded huge sales. A persistent struggle began with the introduction of methods without the use of plows. The soil was cultivated with sowing at depths of 1015 cm with subsoil cultivators that had wide horizontal blades or narrow-tipped plow. This led to the evolution of the soil cultivation process. The key step was to exclude the costly operation of removing the soil. Thus began the struggle to introduce new methods and overcome resistance to change. The technology was gradually refined, and “childhood diseases” (the main concern was the need to remove weeds without plowing) were eliminated gradually by the invention of “smart” herbicides that break down into non-

18

Intelligent Agriculture

harmful elements a little later after their application. In the second half of the twentieth century, agriculture developed rapidly without a plow, especially in the United States and Mexico. This resulted in better growth of the plants. There was development in the conditions of the fields to save energy and preserve the soil. All of these led to improved yields, and reduction of labor and production costs. This is what is now called the “Green Revolution.” • Minimum tillage. The next step in the evolution of tillage was the development of minimum tillage methods, where the affected depth is much lower than in the previous methods. The initial thesis on this method was developed at the end of the nineteenth century. Stairs were designed that could grind the surface layer of the soil along with the vegetation. A kind of mattress was formed that protected the soil from the sun and the cold, retaining moisture and forming a comfortable bed for the seeds. The new technology solved a complex set of problems: it preserved the fertility of the land, it helped to cope with floods and eliminate pests and weeds. • Zero tillage. The minimum tillage method was gradually replaced by zero tillage. This concept has much in common with the ideal final result of the problem-solving, analysis and forecasting tool technique TRIZ (Teoriya Resheniya Izobreatatelskikh Zadatch, which in English refers to the Theory of Inventive Problem Solving). Ideally, zero tillage involves not doing any action on the soil, and at the same time, keeping the soil in an optimal state for the growth and development of plants. Due to the balance among weeds, cultivated plants, and microorganisms included in the biocenosis, the human intervention required is minimal. Currently, zero tillage involves the absence of all tools except the seeding machine.

Consumer Discovery

19

Zero tillage is not “tillage” in the usual sense of the word. It is rather to provide conditions to obtain the optimal structure of the soil. That is, there is no tillage, but the function that the tillage performs is achieved. Thus, it has reached the point where it began. The pattern of evolution is not finished at this stage. Hopefully, new expansions in cultivation technology will become clearer over time. In the meantime, there is a classic spiral. Soil cultivation technology has passed the stage of maximum expansion, and people have seen that it was not as good as they wanted. Now the cultivation of soils is living a period where the soil is minimally affected, but this experience implies a deeper understanding on the subject. In addition, considering the evolution of crop technology, it can be seen that the use of labor is decreasing. Is it really like that? Perhaps the reduction of burden on the worker, the tractors, the seeders, and the soil itself is possible only by replacing physical work with intellectual work. And while the amount of work seems to be constant, a growing part of it moves to the virtual space, that is, to the super system. For example, to control weeds without plowing, it is necessary to invent and manufacture special substances that eliminate weeds that can be harmful to crops and people. In this case, no more hands are required for help, but more heads. Most likely, the evolution of machines and technologies is based on some kind of conservation law and the amount of cumulative work remains constant. That would be something interesting to calculate more accurately.

1.3.2. New Technologies The notable growth of agricultural activity has been associated with the spread of the use of modern machinery for

20

Intelligent Agriculture

different agricultural tasks, as well as improved seeds (including transgenic seeds) and agrochemicals. New planting techniques were also introduced. In all cases, these innovations aim to improve production in terms of quality and yield. The production volume obtained by a cultivated area is called “yield.” • Agricultural machinery: It is used to till the land, cultivate or collect agricultural production. The mechanization of agriculture involved the rampant usage of tractors, seeders, and mechanical harvesters, which made it possible to carry out all kinds of tasks in less time. Agricultural machinery tends to be increasingly complex as it has precision devices and electronic commands, which increases its power, speed, and quality of work. • Inputs: Among the inputs used in agriculture, the ones that stand out for their degree of technological innovation are improved seeds and agrochemicals.  Improved seeds: Improved seeds are those whose genetic material has been modified through the incorporation of information that allows them to acquire a characteristic that they did not previously have, for example: higher yield, better resistance to diseases, greater nutritional volume, and better flavor of fruits and vegetables, among other qualities. One type of improved seeds is hybrid seeds, which arise from crossing plants of different types within the same species, which results in certain desired characteristics to appear combined in their offspring. Another type of improved seeds is transgenic seeds. These are seeds that have been genetically modified through the grafting of genes from other plant or animal species, or from the isolation and modification of their own genes and their reintroduction into the

Consumer Discovery

21

original species. Genetically modified seeds are marketed, in general, by a few, but very large, multinational companies, such as Cargill or Monsanto. The company Monsanto has developed a series of transgenic seeds that have been widely spread in Argentina, especially in the case of soy.  Agrochemicals: Agrochemicals are an essential element in modern agriculture to increase crop yields. Among the agrochemicals are phytosanitary products intended for the protection of crops, thereby ensuring the plants are in good shape and properly develop. Among these are herbicides (control the weeds that invade the crops), insecticides (control the harmful insects), acaricides, fungicides, and bactericides (fight mites, fungi, and bacterial diseases, respectively). Another of the fundamental agrochemical products are fertilizers. These replenish the soil nutrients that the harvest takes away. As in the previous case, the use of fertilizers has increased markedly. • New sowing and planting techniques: Direct sowing and new planting techniques that are in use nowadays in various agricultural fields in all countries, especially in developed countries, are essential to achieve a higher level of food production.  Direct sowing: One of the main innovations of recent times in agriculture is the diffusion of direct seeding as a production system. It is a system in which the seed is grown on a soil that has not been previously plowed, that is, on stubble or residues from the previous crop. The direct seeding system is expanding to other crops, although it is mostly used for soybean cultivation.

22

Intelligent Agriculture

• Remote control agriculture: No literature exists on this new technology. With minimum equipment and the right software, a farmer can listen to one’s cultivation at a distance and gain knowledge on variables such as temperature, humidity, wind speed, nutrients, and pests. Precision agriculture, based on satellite tracking, will allow that and more. However, there is a risk, as the author argues, as the excessive ambition of the transnationals’ could concentrate this technological advance to the detriment of the small farmers of the world. For example, it is a hot and dry night. A farmer is 15 kilometers away from his crop and wonders if it will need irrigation. The farmer cannot afford to spend time and gas, but this technology without having to move from home. Through the Internet, the farmer can receive precise data regarding the weather conditions and remotely activate the irrigation system if the crop needs it. In the ground, there are numerous stakes “separated 10 or 20 meters from each other” with small sensors that record temperature, humidity, direction and wind speed, and other variables. In each, there is a cell phone that transmits the data every 15 minutes, which the farmer can check online. Since the user’s physical location is irrelevant, the farmer could be anywhere in the world: ten or a thousand kilometers from his crop. The planting could be in Oaxaca and the farmer in Stockholm (which makes us wonder if the definition of the word “farmer” will not be stretching too much). One can also program the system so that the planting is automatically irrigated. • Precision farming: Precision agriculture is a system of new technologies and procedures that unite spatial variables of

Consumer Discovery

23

mapping with the realization of appropriate actions of management of the property. It requires the integration of several elementary systems: the global positioning system (GPS), data collection devices (sensors), and GIS, among others. Precision agriculture recognizes the variability within a property and is oriented to do the right thing, in the right way, place, and time. Traditionally, the management of properties is done assuming that each quarter is homogeneous, applying “average” doses. However, it is clear that the existing variability in the properties of the soil influence the capacity of moisture retention, absorption of nutrients, and, therefore, the yield of the crops established on it. In addition, the application of fertilizers and pesticides is usually done uniformly within the barracks, and does not consider the natural variability of the property. This results in the applied doses becoming either excessive or insufficient, with consequent economic and environmental damage.

1.3.3. Advantages and Disadvantages in Technological Advances in Agriculture Recent literature on agricultural and business sciences has various theoretical and empirical studies that establish the existence of multiple advantages and disadvantages regarding the use of new technology for food production. These generated merits and demerits may pertain to the agroindustrial companies or to farmers. Even though several researchers, academics, and professionals of the field have identified more advantages than disadvantages, it is important to establish what would be the main advantages that agricultural producers or agroindustrial companies would have if they use the

Intelligent Agriculture

24

agrotechnology that will be developed, as well as the main disadvantages if they were to do so. This would serve as a point of comparison when deciding whether or not to use the new agrotechnology. The following sections will discuss the main advantages and disadvantages as a result of the new agrotechnology, making a clear separation for each of the essential elements of agricultural and agroindustrial production. (1) Direct sowing ○ Advantages  favors the conservation of soil cover by avoiding or reducing erosion;  improves the use of water, since it maintains the humidity of the soil by covering the soil by a layer of biomass (stubble), which retards the loss of moisture by evaporation;  improves biological activity and increases the content of organic matter in the soil.  improves efficiency in the use of time, since it reduces the amount of work needed; and  reduces the use of machinery (and fuel) and personnel. ○

Disadvantage  In conventional tillage systems, the plow is used, among other things, as a method of weed control. With direct seeding, this method of mechanical control must be replaced by the greater use of herbicides, which means that direct sowing increases the dependence on these phytosanitary products.

Consumer Discovery

25

(2) Agrochemicals ○ Advantages  They allow the production of healthy and abundant food, in an efficient, economical, and sustainable way. Without agrochemicals, the world production of fruits and vegetables, fodder, and fibers would fall between 30% and 40% due to pests and diseases.  An increased production of fruits and vegetables translates into a reduction of costs for the final consumer and therefore greater access to food by the population.  Agrochemicals are a tool to control pests that directly affect the health of people, such as malaria.  The control of termites and other insects that can put at risk houses and other constructions are also controlled by agrochemicals.  Agrochemicals are the result of rigorous research and development (R&D) processes, which last an average of nine years and have an average cost of 200 million dollars.  The R&D industry constantly invests in the search for new technologies and scientific advances.  In this search for new knowledge, the industry supports the talent of scientists from different disciplines and highly specialized professionals.  The R&D industry permanently transfers technology through products and services in favor of sustainable agricultural development.

Intelligent Agriculture

26

 The latest technological developments allow farmers to produce more in less space, as well as to have products that degrade easily and leave a lesser impact on the environment. ○

Disadvantages  An exhaustive training of the rural worker and producer must be carried out to achieve an effective manipulation and use of the agrochemicals. The ignorance of efficient cultural practices, in many cases, results in rivers and streams being contaminated, thereby causing health problems in people and animals, and affecting the environment in general. In Latin America, 865,000 tons of industrial waste is generated daily, of which 15,500 tons are hazardous. Unfortunately, less than 10% of the hazardous waste is treated properly, with the remaining 90% discharged into streams and garbage dumps.  Misuse can cause serious illness or death, contamination of soil and water, damage to livestock and wildlife, and reduction or elimination of the natural enemies of pests. For this reason, it is essential that proper management and supervision of the use of pesticides is undertaken.  In general, it is common to recognize the problem of pesticides in the case of agricultural projects that aim to produce crops; however, they are often overlooked when they are used to reduce losses, especially after harvest. There are several types of pests (including fungi) that cause losses

Consumer Discovery

27

during storage. In developing countries, the problem is further complicated by weather conditions and the lack of adequate storage infrastructure. Usually, products that are stored in bulk are fumigated or treated with diluted insecticide powders. Fumigation can be extremely dangerous for humans, requires specialized equipment and training, and can leave toxic residues in food. (3) Improved and transgenic seeds ○ Advantages  It is assumed that transgenic products are resistant to pests, last longer without withering, and have a better appearance.  With these products, there will be more food in the world and therefore less hunger.  However, those who are making these products are the same companies that manufacture agrochemicals: poisons, fertilizers, and improved seeds. These big companies are Monsanto and DuPont.  Bayer and Pioneer are also producing transgenic seeds but these companies appear under other names because they are linked to other companies. They are offering transgenic seeds that resist pests and herbicides, which gives higher production and better quality.  They assure that when cultivating these seeds, fewer agrochemicals are needed and therefore they claim that it results in less risk of contamination for the people and the environment. But these are the very

Intelligent Agriculture

28

companies that 40 years ago offered agrochemicals with many advantages to improve production, eliminate pests, and have more crops.  The result after years of use is that pests are becoming more resistant, and every year more people die because of intoxication or become sick by using these products; soils and water are contaminated; and producers, year after year, have to buy the improved seed, thus losing the wisdom regarding the production of their own seeds. ○

Disadvantages  The emergence of transgenic agriculture opened a wide debate because it has not been demonstrated with certainty that transgenic vegetables harm their consumers in any way. Those who oppose argue that it has not been proven otherwise.  For now, it is not known how the consumption of these products will affect the health of people.  Several scientists say that they can cause cancer, develop allergies, and result in poisoning, organ damage, and antibiotic resistance; that is, they will no longer cure the person.

1.4. INTELLECTUAL PROPERTY ANALYSIS Technology is the backbone of a digital economy and much of its value is found in computer programs. Indeed, the growth of all economic sectors is increasingly dependent on computer programs. This has important implications for

Consumer Discovery

29

intellectual property legislation. Until the end of the twentieth century, the functionality of most innovative products, particularly those based on semiconductors, resided mainly in computer equipment. It was clear that they had to be patented. Nowadays, the continuous sophistication of design tools and semiconductor technology means that physical objects are no longer the only basis for innovation. That is, the technical functionality is progressively moving from physical media to software. In spite of this, in many places, the legislation does not consider inventions that contain software capable of being protected by patents, or are only protected in a very limited way. The enormous economic growth and innovative potential of technological companies that produce products that combine physical media with computer programs, and that of the software sector in general, indicate that the time has come to reconsider intellectual property legislation and to adapt it to the commercial reality of today. The great advantage of software is that engineers and designers can easily develop products equipped with new technical features, put them on the market or grant licenses to third parties to market them, and correct errors and distribute new programs through a simple online update. In many cases, an invention is more quickly set up in a computer program than in a physical medium, which is a more profitable way of getting a product to the market. Consumers can enjoy affordable and continuous access to the latest advances. In addition, the relatively low capital investment needed to create software solutions makes it easier for small businesses and emerging companies to access the market. However, these companies need to effectively protect their intellectual property to obtain reasonable benefits from their investment in R&D.

30

Intelligent Agriculture

1.4.1. What Intellectual Property Rights Are Important to Protect Computer Programs? In the past, intellectual property laws helped the software sector succeed by offering its programmers the chance to recover at least part of the value of their innovations in the market. Since the 1960s, at least, the software sector has relied on three different regimes for the protection of intellectual property: trade secrets, copyright, and patent law. The scope of protection offered by each of them has varied considerably over time, as has the sector’s dependence on these regimes. History shows that patent law is the most effective framework for protecting inventions. Despite this, in many countries, a distinction is made between inventions embodied in physical media, which can be protected by patents, and inventions implemented by computer programs, which are protected by copyright. In a world in which the Internet (and not physical media, such as CDs) is the main channel for the distribution of computer programs, this legal distinction makes the creators of inventions containing software have difficulties protecting and effectively promoting their commercial value through intellectual property systems. These innovations contribute to society and are no less important than those based on physical supports. Computer programs, including inventions containing software, are products in their own right, regardless of how they are distributed. Would it not then be reasonable for these inventions to be protected effectively by patent laws? Currently, many technological innovations depend on advances in computer programs, e.g., those that use such programs and that have revolutionized the smartphone. Between 2009 and 2013, the total lines of code added in the chips (the brain of the smartphone) billed by Qualcomm went from 330 million to

Consumer Discovery

31

3,300 million. This spectacular, unprecedented evolution was the result of years of high-risk investment in R&D. The functions implemented through computer programs are making many of the everyday products safer and more efficient, thereby generating higher-quality results. This is creating entirely new possibilities and capabilities, such as smart grids, digital manufacturing, real-time farm management systems, smart cities driven by interconnected platforms (Internet of Things), and digital healthcare. It is estimated that the digital economy, which is based mainly on innovations assisted by computer programs, already represents 22.5% of the world economy. The global R&D spending on computer programs has also grown rapidly and has gone from 86,000 million US dollars in 2010 to 142,000 million in 2015: an increase of 65%. The United States has one of the most active software sectors in the world (Shapiro, 2014). It is estimated that in 2014 the direct contribution of the sector to the GDP of the country was 475,300 million (1.07 billion indirectly). In addition, 2.5 million direct jobs were created, and 9.8 million indirect jobs. 1.4.1.1. The Benefits of Patent Protection As a general rule, any new invention in a field of technology can be protected by a patent if the invention is novel, not obvious and useful (patentability criteria are established in national patent laws). Patent protection means the following important benefits: • offers a reasonable return on the commercial success of an invention; • facilitates small businesses and start-ups, whose activity is innovation, and the establishment of fruitful business collaborations;

32

Intelligent Agriculture

• promotes the systematic dissemination of knowledge from the publication of the patent, which is an important driver for innovation; and • helps attract investment partners and supports business expansion. However, laws usually do not treat inventions that contain software in the same way as other innovative technological advances. The reason may be that the nature of the innovation or the protection obtained with the different intellectual property rights is not understood. 1.4.1.2. Response to Criticisms of Software Patents According to certain critics, the expense in R&D associated with the development of inventions containing software is not the same as that associated with the inventions of other technological fields. The truth is that many of these innovations, such as systems to improve energy efficiency, advanced medical diagnostic tools, security solutions for smart cars, and surgical robots, take years of research, development, and commercialization. In other cases, it is said that software patents are of low quality or that what they do in the end is grant protection to “mathematical formulas,” and that copyright and trade secrets already provide adequate protection to the intellectual property of these programs. Although the advantage of copyright is that protection is automatic and free whenever the work is original, relying on it as the only safeguard helps to protect only the source code or the object code against its literal copy, but does not protect the underlying invention created for the computer program. Trade secrets also do not require formal registration beyond confidentiality agreements, and their protection is one of the least developed areas of intellectual property law. Even in

Consumer Discovery

33

countries that have enacted laws in this regard, it is not possible to protect innovations since the public can easily find their secret, either by independent discovery or reverse engineering. In addition, trade secrets should not be protected in the case of standardized technologies that promote interoperability, such as smartphone communication technology, since organizations that set technical standards need to exchange technical information in a non-confidential manner. Trade secret protection does not allow this type of information exchange. Thus, copyright and trade secrets are complementary forms of protection but neither provide the same benefits as patents nor the same incentives to invest in the underlying innovation. The definitive criterion for patent protection should be the quality of the invention and not its mode of application. The decision to put an invention into service using hardware or software is often a design issue that should be left to technical experts instead of being limited by patent law. Distinguishing between inventions containing software and inventions that do not contain software to justify discriminatory treatment violates the purpose of patent law and could hinder technological progress. Quality should not be a concern because the patent examination process is designed to offer legal protection only to inventions that meet certain strict criteria. Future inventors should present a novel idea  useful and not obvious to an expert in the field. Patent examiners are trained to judge whether the proposed invention represents a technological advance. The important thing would be for the examiners to have the right tools to carry out that evaluation instead of excluding inventions containing software from patent protection.

34

Intelligent Agriculture

1.5. VALUE PROPOSAL In order to boost production, strengthen food security, and improve the income of farmers and small producers, the Ministry of Agriculture, Livestock, Rural Development, Fisheries and Food (Sagarpa) launched the Proagro Productivo. According to data obtained from the Second Government Report, in 2015, agricultural production grew by 6.5%, with an increase in the production of cane and sugar and a growth in the volume of corn, beans, and wheat. Also, in the first seven months of 2012, 11,384 million pesos were given as incentives before planting, with the purpose that two million producers, who cultivate 10.6 million hectares, will have the resources to acquire improved seeds, fertilizers, equipment, and supplies. Another important information that Sagarpa reveals is that, at the end of the first half of 2015, the agricultural and fishing trade balance registered a surplus of 535 million dollars, which represented a 39% reduction in the deficit  the largest in the last eight years for a similar period. Last year’s production with respect to grain and oilseed crops (rice, corn, beans, wheat, soybean, and sorghum) covered 68% of national consumption, which represented a growth of five percentage points compared to 2012. “The goal, is to achieve that 75% of the internal consumption of these products occur in the country,” argues Sagarpa in the statement. A differentiated aspect between Proagro Productivo and Procampo is that the current program handles differentiated stimuli. Producers who plant for self-consumption are granted an incentive of 1,500 pesos per hectare  a figure that represents 200 pesos more in real terms than what they received in 2013. Transitional and commercial producers receive 537 pesos more per hectare. Sagarpa states that, thanks to the

Consumer Discovery

35

increase in stimuli, it has been possible for producers to have timely resources for the acquisition of fertilizers, correctors, substrates, seeds, and vegetative material, as well as phytosanitary production, lease payments, administrative expenses, and marketing and rights. The most important sowings of each state are described in the forthcoming text: Aguascalientes. It is the second largest producer of guava, with 28.7% of the national sown area and 15% of the total production and the third largest producer of irrigated fodder maize, covering 12.6% of the planted area of the country and 11.9% in national production. Baja California Sur. It is the fifth largest producer of red tomatoes by irrigation, with 6.1% of the national sown surface and 6% of the total production. It is the fourth largest producer of asparagus by irrigation, covering 9.7% of the total area and 4% of the total production, and is the eighth largest producer of green chili by irrigation, with 1.4% of the total national surface and 2.4% of the total production. Baja California. It is the second largest national producer of onion by irrigation, covering 15.8% of the sown area and 13.6% of the total production. It is also the second national strawberry producer, covering 32% of the planted area and 46.6% of the national production. With regard to olives and dates, it is the leading national producer, covering 70.5% of the planted area and 42% of the total national production for olives and 22% of the total sown area and 41.3% of the total production for dates. Campeche. It is the foremost producer of temperate rice, with 24% of the total national production value.

36

Intelligent Agriculture

Chihuahua. It is the leading producer of irrigated cotton, covering 56.7% of the national sown area and contributing 46.6% to the national production, and also for forage oats of the rainforest, covering 35.1% of the total sown area and 40.9% of the national production. In apple production too, it is the first, covering 43.1% of the national sown area and 73.7% of the total production. It is the foremost nut producer, contributing to 60.7% of the national sown area and 56.2% of the national production. Chiapas. It is leading producer of cherry coffee, with 32% of the sown area and 39% of the national production. It is the second largest banana and payaya producer, covering 26% of the sown area and 28% of the national production and 10% of the sown area and 16% of the national production, respectively. It is also the second largest producer of irrigated tobacco, covering 5% of the sown area and 6% of the national production. Coahuila. It is the leading producer of irrigated forage oats and watermelon, covering 17.4% of the national sown area and 19.7% of the total production value and 21.8% of the total area and 19.1% of the total national value, respectively. Colima. It is the foremost producer of lemon, covering 20.3% of the planted area and 30.5% value of the national total production in 2007. It is the fourth largest banana producer, with 7.3% of production and 6.7% of the national production. Estado de México. It is the leading producer of temporary potatoes, with 35.8% of the national production and 25.2% of the total sown area. As the leading and only producer of carnation, it covers 100% of the national production and 100% of the planted area. It is the second

Consumer Discovery

37

largest producer of seasonal grain corn, forage oats, and peach, covering 11% of the national production and 7% of the total sown area, 7% of the planted area and 19% of the national production, and 7% of the sown area and 16% of the national production, respectively. Durango. It is the leading producer of irrigated fodder maize, contributing to 21% of the national production value and 22% of the total sown area. Moreover, it is the second largest producer of beans and apple, covering 35.2% of the national production value and 36.8% of the total sown area and 6.8% of the national production value and 17% of the total sown area, respectively. Guanajuato. It is the foremost producer of irrigated sorghum and barley, covering 38.2% of the national production and 25.6% of the total sown area and 90% of the national production and 89% of the total sown area, respectively. It also leads in pomegranate and strawberry (perennial) production, with 32% and 58% of national production, respectively. Being the second largest producer of irrigated corn and wheat, it covers 8% of the planted area and 9% of the national production and 15% of the sown area and 16% of the total production, respectively. It is the second largest asparagus and seasonal sorghum producer, covering 23% of the planted area and 27% of the national production and 8% of the national production and 7% of the total sown area, respectively. Guerrero. It is the leading producer of copra, covering 61.7% of the planted area at the national level and contributing to 71.9% of the national production. It holds the second place in the production of mango, with an area sown equivalent to 12.2% of the national total, contributing to 18.1% of the national production. It is the

38

Intelligent Agriculture

third leading producer of irrigated melon, contributing to 17.3% of the national sown area and 13% of the total production. Hidalgo. It holds the first place in barley production, covering 43.1% of the national sown area and 41.4% of the total production. As the leading producer of maguey pulquero, it covers 47% of the planted area and 79% of the national production. Fifth place in bean production, it covers a planted area equivalent to 4.6% of the national total and contributing to 6% of the national production. It is the foremost producer of green alfalfa, covering 4% of the total sown area and 3% of the national production. Jalisco. It is the foremost national producer of seasonal grain corn, covering 7.7% of the surface area and 13.8% of the total production, and the second largest national producer of sugarcane, covering 9.6% of the planted area and 11% of the production. It is also the second largest national producer of irrigated watermelon, covering 9.3% of the sown area and 15.6% of the production. It holds the first place in the national production of seasonal fodder maize, covering 29.8% of the sown area and 26.9% of the total production. Morelos. It holds the first place in the production of seasonal tomato, representing 24.3% of the national sown area and 31.1% of the country’s production for this crop. It is the second largest producer of avocado and nopales, with 2.1% of the total planted area and 2.3% of the production and 21.8% of the total sown area and 23.7% of the national production, respectively. It is the sixth largest producer of irrigated rice paddy, covering 4% of the national planted area, representing 8.1% of the value of the total production.

Consumer Discovery

39

Nayarit. It is the leading producer of irrigated tobacco and irrigated rice paddy, with 92% of the national production and 93% of the total sown area and 22% of the planted area and 19% of the national production, respectively. Nuevo León. It is the third largest producer of seasonal forage sorghum, covering 7.7% of the national production and 10.6% of the total sown area. It is also the fourth largest producer of irrigated potatoes and oranges, with 14.7% of the national production and 8.7% of the total sown area, and 10.7% of the national production and 7.7% of the national sown area, respectively. Oaxaca. It is the leading producer of seasonal melon, with 65.4% of the national production and 57.3% of the national sown area and the third largest producer of sugarcane, with 7.7% of the total production and 7.7% of the planted area. Puebla. It holds the first place in the production of gladiola (thick) by irrigation, covering 35.2% of the sown area and 34.6% of the national production. It is the third largest producer of cherry coffee, apple, and irrigated carrots, with 18% of the national production and 9% of the total sown area, 6% of the national production and 13% of the total sown area, and 23% of the total planted area and 20% of the national production. Querétaro. It is the sixth largest producer of irrigated forage corn, with 10% of the national production and 9% of the total planted area. It is also the seventh largest producer of sorghum irrigated grain, with 2% of the national production and 1% of the total sown area. Quintana Roo. It is the sixth largest producer of green chili from the rainforest, covering 11% of the area sown

40

Intelligent Agriculture

and 6% of the national production, eleventh largest producer of papaya, covering 1% of the sown area and national production, and the twelfth largest producer of oranges and sugarcane, covering 1% of the national sown area and 1.1% of the national production for oranges and 3% of the national production value and 4% of the total sown area for sugarcane. San Luis Potosí. It is the second largest producer of irrigated corn, covering 21% of the national production and 18% of the total planted area, the third largest producer of oranges, covering 13% of the sown area and 10% of the national production, and the fourth largest producer of sugarcane and green chili by irrigation, covering 7% of the national production and 9% of the total sown area for sugarcane and 10% of the national planted area and 6% of the total production for green chili. Sinaloa. It holds the first place in the national production of corn, irrigated grain variety, red tomato by irrigation, green chili by irrigation, and irrigated potatoes, covering 35.4% of the planted area and 49.2% of the national production, 34.3% of the sown area and 37.3% of the national production, 13.7% of the planted area and 32.3% of the national production, and 31.5% of the national sown area and 27% of the total production, respectively. Sonora. It holds the first place in wheat production for the irrigated grain variety, asparagus, and grape, covering 52% of the national production and 50% of the total planted area, 55% of the national production and 48% of the total sown area, and 87% of the national production value and 71% of the total sown area.

Consumer Discovery

41

Tabasco. It is the second largest national producer of seasonal paddy rice, covering 33% of the sown area and 31% of the total production. It is also the leading national banana producer with 18% of the sown area and 32.1% of the national production and the third largest national producer of pineapple with 4% of the national sown area and 5% of the total production. Tamaulipas. It holds the first place in the national production of seasonal grain sorghum, covering 50% of the planted area and 47% of the national production, and third place in the national production of onion by irrigation, with 11% of the sown area and 13% of the national production. It is second largest national producer of orange, mandarin, and grain sorghum, with 10% of the planted area and 11% of the national production, 20% of the planted area and 25% of the national production, and 44% of the sown area and 33% of the national production, respectively. Tlaxcala. It is the second largest producer of seasonal barley, covering 13.4% of the total national planted area and contributing to 19.1% of the national total production. It is the leading producer of temporary wheat, contributing to 33% of the national production. In 2007, Tlaxcala stood out as the second largest producer of maguey pulquero (perennial) and the third largest producer of irrigated spinach nationwide. Veracruz. It holds the second place in the national level in the production of cherry coffee and seasonal watermelon. For cherry coffee, Veracruz contributes to 32% of the planted area and 39% of the national production, whereas in terms of seasonal watermelon, it contributes to 21.6% of the national sown area and

42

Intelligent Agriculture

22.7% of the national production. Veracruz is also the leading producer of sugarcane, green chili for the rainforest, and papaya, with more than 259,000 hectares sown and covering 36% of the national production for sugarcane, 26% for green chilli and 43.5% of the national sown area and 43.3% of the national production for papaya. Yucatán. It is the fourth leading producer of papaya and irrigated cucumber, with 6% of the total planted area and national production for the first and 5% of the national production and 3% of the total sown area for the second. It holds the sixth place in lemon production, with 3% of the sown area and 4% of the national production, and seventh place in orange production, with 4% of the national sown area and 3.7% of the national production. Zacatecas. It is the leading producer of beans with 35% of the national production value and 37% of the total sown area and the second largest producer of irrigated beans, grape, and seasonal grain wheat, with 12% of the national sown area and 13.5% of the total production, 13% of the sown area and 10% of the national production, and 19% of the national production and 29% of the area sown, respectively. Once the main crops have been studied, a list is created with data from the Tax Administration Secretary (SAT). The list contains the legal entities that registered the planting and production of broccoli in the country as part of its commercial activity. Due to data protection laws, it is not possible to obtain data from individuals and their activities.

Consumer Discovery

43

Table 2. The Most Important Sowings in Mexico. Total Production (Thousands of Tons)

2014

2015

Annual Value (%)

Irrigation Grain sorghum

1,093.4

1,136.6

4

Grain corn

935

1,044.6

117

Grain oats

592.2

584.6

1.3

Barley

237

264.5

11.6

Tomato Forage oats

40.3

41.5

3

173.9

167.6

3.6

25.2

24.6

2.4

244.6

248.8

1.7

Grain sorghum

513.6

403.4

21.5

Grain corn

612.2

429

29.9

Forage oats

85

125

47.1

Bean

43.9

51.8

Green tomato Forage corn Temporary

Grain corn

18

1,398.5

1,581.4

13.1

Potatoes

120.4

141.9

17.9

Grain rice

34.4

37.1

7.8

Grain oats

62.8

63.7

1.4

Grain oats

17.2

19.1

11

4,195.5

4,112.8

2

14.6

15.4

Perennial Green alfalfa Asparagus

5.5

Nopales

5.40

5.41

0.2

Guava

1.1

1.3

18.2

Peach Carnation (thousands of dozens) Avocado

32.2

33.7

4.7

4,802.9

5,274.3

9.8

0.7

1

42.9

Intelligent Agriculture

44

Table 3. Broccoli Producers in Mexico. (1) Química Farmacéutica Esteroidal

(2) Química Farmacéutica Mexiquense

Cerrada 15 de Septiembre No. 40 Col. Francisco Villa San Juan

Cerrada 15 de Septiembre No. 80

Ixtayopan

Col. Francisco Villa San Juan

Del. Tláhuac

Ixtayopan

13520, México, D.F.

Del. Tláhuac

México

13520, México, D.F. México

(3) Agrícola Cueto Produce

(4) Agrícola la Minita

Herrera y Cairo No. 5

Carretera Panamericana

Col. Sayula Centro

Km. 291

49300, Sayula, Jalisco

Col. la Fortaleza

México

38300, Cortázar, Guanajuato México

(5) Agrícola las Montañas

(6) Agrícola Nieto

Calzada CETYS No. 2799, EAL3

Hacienda San Antonio S/N

Fraccionamiento Rivera 21259, Mexicali, Baja California

Fracción Labradores Villagrán

México

38260, Villagrán, Guanajuato México

(7) Agrícola Río Fuerte

(8) Agriexport

Zaragoza No. 144 Norte Edif. San

Nainari No. 1089 Oriente

Isidro Interior 21

Colonia Centro

Colonia Las Fuentes

85000, Cajeme, Sonora

81200, Ahome, Sinaloa

México

México (9) Tierra Viviente Agricultura Orgánica

(10) Agropecuaria Sanfandila Carretera Lagos León Km. 1

Isaac Arriaga No. 142

Colonia Cañada de Ricos

Colonia Centro

47450, Lagos de Moreno,

60000, Uruapan, Michoacán

Jalisco

México

México

Consumer Discovery

45

Table 3. (Continued ) (11) América Alimentos

(12) Comercializadora Gab

Avenida Santa Ana Tepatitlán 90 A

Carretera Panamericana Km. 5

Colonia Agrícola 45236, Zapopan, Jalisco

Colonia Rancho Grande 36543, Irapuato, Guanajuato

México

México

(13) Congelados Don José

(14) Exportadora San Antonio

Carretera León-Cuerámaro Km. 12.5

Rancho San Antonio Fraccionamiento Los Órganos

Colonia San Cristóbal

S/N

37230, Silao, Guanajuato

38260, Villagrán, Guanajuato

México

México

(15) Grupo Altex

(16) Las Cinco Estaciones

Paseo de las Palmas No. 820 1er.

Morelos No. 309

Piso

Colonia Centro

Lomas de Chapultepec

38300, Cortázar, Guanajuato

11000, México, D.F.

México

México (17) Mar Bran

(18) Megafrescos del Bajío

Boulevard San Roque No. 1540

Carretera SalamancaLa

Colonia Centro

Ordeña Km. 6

36590, Irapuato, Guanajuato

Colonia Xoconoxtle de Arriba

México

36861, Salamanca, Guanajuato México

(19) Mexicana de Nutrimentos

(20) North Countryside

Niño Artillero No. 450-1 Colonia Universitaria

Selections de México Abasolo No. 202

78290, San Luis Potosí, San Luis

Colonia Emiliano Zapata

Potosí

64390, Monterrey, Nuevo León

México

México

(21) Productos Selectos Marroko

(22) Rijk Zwaan de México

Avenida Emprendedores No. 109

Calle de las Gladiolas No. 4176

Colonia Ciudad Allende

Colonia Hacienda Molino de

67350, Allende, Nuevo León

Flores

México

80155, Culiacán, Sinaloa México

Intelligent Agriculture

46

Table 3. (Continued ) (23) Sun and Green Produce

(24) Tecno Agro Industrial

Avenida Vallarta No. 4906

Km. 18 Ejido Islita San Luis Río

Colonia Vallarta La Patria 45020, Zapopan, Jalisco

Colorado Colonia Ejido Islita

México

83500, San Luis Río Colorado, Sonora México

1.5.1. Broccoli Producers in Mexico In Mexico there are relatively few farmers and agroindustries that are dedicated to growing broccoli; however, in the last two decades, there has been an increase in production, particularly in the region of the shoal and north of the country, and farmers and agroindustries are significantly increasing the amount of land for its cultivation due to the strong demand that has been noted in these last two decades, both in the Mexican market and in international markets. The production of broccoli both in Mexico and in any other part of the world entails too many risks, since the ripening time, the quality of the cultivated land, and the irrigation water are directly affected by the variations in the weather throughout the period of production, particularly the low temperatures that are registered in agricultural fields, as is the case in the region of the shoal and the north of the country. However, the new varieties of broccoli, and/or the improvement of existing varieties in the market, are allowing farmers to achieve a greater crop yield even when low temperatures are recorded. This is due to the genotype of the plant and the proper selection of broccoli variety for the type of cropland that is intended to be used. Therefore, the

Consumer Discovery

47

development of new procedures and techniques for growing broccoli combined with the use of more sophisticated machinery and technology in the field are allowing an increase in the percentage of farmland that is being devoted to the production of broccoli. In this sense, Table 3 presents the main agricultural and agroindustrial producers and traders that produce broccoli throughout the national territory, which has been obtained from the main databases of national agricultural groups.

This page intentionally left blank

CHAPTER 2 SYSTEM ARCHITECTURE

2.1. INTRODUCTION The objectives and goals of agricultural production are too similar to the objectives and goals of industrial production because both require extensive periods for the development of their level of production, generally using models dependent on non-linear production control and prediction that require various restrictions for its application (Ding et al., 2018). However, it is in the field of agricultural and agroindustrial production that several researchers, academics, and professionals, with diverse areas of knowledge, are currently orienting their theoretical and empirical studies more and more. These studies usually pertain to the development and application of the models of control and prediction of production, especially in irrigation systems, in the efficient use of agricultural machinery, in food production systems, in the processing of products, and in their storage. In the same vein, it is essential for plants, especially for broccoli, to receive efficient irrigation during their growth and development phase. Irrigation is usually more efficient when the irrigation systems are based on technology, 49

50

Intelligent Agriculture

for example, robust control systems, optimal control systems, predictive control systems, or non-linear control systems (Salahou, 2013). However, normal controls and robust controls do not provide an accurate description of irrigation control and prediction, and so plant irrigation generally involves non-linear models, with long periods for execution and various restrictions. This causes serious difficulties while creating an exact model that is completely error-free. Therefore, to improve the accuracy of the model, artificial intelligence controls are currently used, which are fully integrated into various devices or sensors that can measure and control the level of irrigation of the plants. Artificial intelligence controls are usually integrated by complicated mathematical models that can seamlessly solve any type of problem that occurs in the production system of companies, including overcoming the various restrictions associated with the production channel (Wu, 2008). Therefore, control and prediction models with artificial intelligence, which are usually integrated into computer hardware, are considered by various researchers, academics, and professionals in the field of basic and economic sciences as fundamental electronic devices for improving the control and prediction of irrigation systems in agricultural fields (Delgoda, 2016). The fundamental objective of these systems involves controlling the water level that must flow through the channels. This allows irrigating the plants with only a certain percentage of water at a time, as in the case of broccoli, where such irrigation would yield a higher level of production. For cotton, McCarthy (2014) used an irrigation model based on a strategy that generated exceptional simulation of the agricultural field and its irrigating system. Likewise, Aguilar (2016) demonstrated that predictive control of irrigation yielded better output than the absence of control in a

System Architecture

51

prediction system. Therefore, it is better to apply and manage a system of control and prediction of the irrigation system in an agricultural field. Control is through automated processes by means of computers (hardware). Additionally, Roca (2016) developed a hierarchical control model to improve the irrigation volume in an agricultural plantation, achieving significant results that allowed the company not only to improve the irrigation system of the plantation but also to save water. Similarly, in the case of irrigation systems of large agricultural fields, the use of sensors in hardware systems, supported by software and computer systems, is more effective and efficient (Ding et al., 2018). In this same order of ideas, Negenborn (2009) proposed a new method for the control of irrigation channels, using a time series model based on a noninteractive cascade predictive prediction and control, which gave good results. He also pointed out that non-interactive methods commonly generated a higher level of performance than interactive methods, especially when the time series data were greater than 30. Therefore, the data and information associated with the interactive algorithms and the control and prediction models of agricultural and agroindustrial production help generate better results than the models associated with the non-interactive algorithms (Scattolini, 2009). Also, Farhadi and Khodabandehlou (2016) proposed a production control and prediction model with a two-level architecture, finding that the second level generates a better distribution of the production control and prediction model because it has a higher level of productive performance. In addition, Figueiredo (2013) applied a control and prediction model of automated production supported by hardware to generate a more efficient and effective irrigation channel, which is managed through supervision control and data acquisition. Therefore, this type of automation of both irrigation systems and the entire production process of agricultural and

52

Intelligent Agriculture

agroindustrial production generates information not only more robust but also more reliable, that facilitating the control and prediction of agricultural and agroindustrial production. Finally, Horváth (2015) concluded that the automation of control models and production prediction generates better results because it achieves continuous interaction between the various irrigation channels and production processes. Horváth (2015) also developed and implemented an additional component that can be perfectly applied through a production control and prediction model. This component helps eliminate, as much as possible, the errors that predictive control could generate, both at the level of utilization of the water as irrigation systems of agricultural fields, thereby showing results that were not only effective and efficient but also significantly improved. Therefore, various electronic devices, which are generally transformed into hardware components, efficiently and effectively control systems and predict agricultural and agroindustrial production.

2.2. WIRELESS SENSOR NETWORK The wireless sensor network (WSN) is practically a communication network formed by a certain number of sensor nodes, which is generally used for the real-time monitoring of a series of determined information (for example, the intensity of solar radiation, temperature, humidity, wind speed, evapotranspiration, the amount of accumulated rain, and the concentration of greenhouse gases) and roam a remote area where the nodes were strategically placed. The information is usually processed into a series of data that are sent for analysis to various researchers (Tian et al., 2013). The data and information obtained from a series of variables through the different measurement techniques of WSN are often used for

System Architecture

53

decision-making processes not only in the business sector but also in relation to various social issues, such as the usage or not of personal vehicles to try to reduce the poor quality of air that is breathed in large cities, or the avoidance of exercise outdoors. Moreover, with the rapid development of information and communication technologies that directly affect our lives, technology integration in virtually all social and economic activities of human beings, technology usage in sensors, and the constant use of microsensors and nodes in our daily activities (Akyildiz et al., 2002), the WSN is becoming one of the most attractive constructs for its analysis and development. WSN is increasingly attracting the attention of researchers, academics, and professionals from all fields of scientific knowledge, primarily because it has application in all fields of industry, for example, in the construction of smart homes, in medical and health services, and in financial services, among others (Werner-Allen et al., 2006). Currently, WSN is becoming one of the most popular constructs in which researchers are orienting their studies, because the development of any technology in society has a significant impact in general. The WSN usually contains a series of sensor nodes, a receiver node, and one or more information reception terminals, but when there is a large number of sensor nodes that are practically scattered in one or more remote areas of a given agricultural field, all these sensor nodes should have a network that is fully adapted to work properly, or else there would be too many problems with its communication system (Tian et al., 2013). Therefore, in such a scenario, a secure environment will be required to collect all the information through the sensor nodes, which will be communicated basically by the network developed exclusively for it with the receiving node. This receiving node will send the information

Intelligent Agriculture

54

to the terminal or terminals indicated through a global positioning system (GPS) or via a satellite for analysis and thereafter added to a database, which can be fully configured and managed by the user or users according to their needs for presentation and analysis of the information (Fengyuan et al., 2003). Figure 1 exemplifies the aforementioned in a better manner, with a structure that has a WSN. Building a WSN system requires the development and integration of many hardware and software components (Aqeelur-Rehman et al., 2018). The system includes a base station and a distributed number of sensor nodes. Each sensor node is a combination of sensors, a microcontroller (µC), and a XBee module. Data collected from the sensors are sampled by a user application program on each sensor node, which is communicated to the base station (Aqeel-ur-Rehman et al., 2018; Methley et al., 2008). This study focuses on developing a sensor node using Arduino, XBee modules, and low-power Raspberry Pi as the base station to transfer all data to the computer. Arduino is a widely used open-source single-board Figure 1. Structure of WSN.

Sink node Internet/Satellite

Remote area Data terminal users

Sensor nodes

System Architecture

55

microcontroller development platform with flexible, easy-touse hardware and software components. The construction of the WSN will allow agricultural and agroindustrial producers to gather important information on broccoli production, such as, for example, plant growth, flowering, leaf size, growth of the button, as well as other soil and meteorological variables that will be used for the control and prediction of agricultural production. Information is obtained with the help of sensors placed in strategic locations of the agricultural field. These sensors will send information from time to time through the sensor nodes, thereby enabling the user to make better decisions for the management of production. Figure 2 shows the overall system architecture of the WSN for smart farming, as proposed and developed in this project. As can be seen in Figure 2, the architecture of the WSN in the project is simple and basic, thus practically concentrating only on the functionality to obtain the required information to generate a control model. It also generates a prediction of the production in such a way that its structure allows a significant growth with the use of some other sensors that are considered pertinent. In addition, in a medium term, the architecture of the WSN also incorporates information and additional data for the best business decision-making. Therefore, initially, the architecture and structure of the WSN will be considered for the prediction of broccoli production among the producers of the state of Aguascalientes (Mexico). Later, it will be possible to scale the system according to the information needs once it is in operation in the agricultural fields.

2.2.1. Sensors The United Nations has estimated that the world population will go from the currently existing 7.4 billion people to just

56

Figure 2. Overall System Architecture of WSN. Raspberry Pi Base Station

PC/Laptop Data Display on PC

LABVIEW GUI Xbee Coordinator IEEE 802.15.4

User

SQL Database

Data transferred wirelessly

XBee Router 1

XBee Router 2

XBee Router N

IEEE 802.15.4

IEEE 802.15.4

Arduino Mega 2560

Arduino Mega 2560

Sensors

Sensors

Sensors

Node 1

Node 2

Node N

Intelligent Agriculture

IEEE 802.15.4 Arduino Mega 2560

System Architecture

57

over 9.55 billion people by the year 2050, and even that it may have a population of 11.2 billion people for the year 2100 (Bogue, 2017). In fact, an exponential population increase is not the main problem facing the nations of the world, but rather the demand for food associated with this growth, which is causing drastic, irreversible changes in the environment. Such an outcome is primarily because of the implementation of various conventional methods in the production of agricultural and agroindustrial products in most developing countries and emerging economies, as is the case of Mexico. This is why it is essential that such countries change drastically to incorporate new production methods that are directly associated with the use of technology (Bogue, 2017). Under this scenario, it is required that researchers, academics, and professionals from all fields of scientific knowledge, associated with the agricultural and agroindustrial producers and the public administration of the three levels of government, work in coordination to carry out the various significant changes that the technification of the global agricultural field entails, mainly that of the developing countries. These changes include, for example, the scarcity of labor, the unpredictable variability of the climate, the reduction of fertility of the arable land, the erosion of the same farmland, the poor management of agricultural production, the existing pressure on agricultural producers to sell their land for the urbanization of cities, and low expectations of the governments for the application of more and better laws that promote greater sustainability practices, such as the reduction of drinking water and agrochemicals in food production (Bogue, 2017). In this sense, precision in agricultural production systems is one of the central elements in the development of agrotechnology, in which most of the theoretical and empirical studies

58

Intelligent Agriculture

of researchers, academics, and professionals are focusing, in addition to the fact that it is one of the most important concepts of agricultural management that is generally based on the observation, measurement, and response of the existing variability to the internal and external interiors of the agricultural field (Bogue, 2017). More specifically, Blackmore (1994) concluded that the use of sensors in the development of agrotechnology is essential because they are practically designed to optimize agricultural production by carefully adapting the cultivated land and the management of agricultural crops, with the sole objective of maintaining both an optimal quality of the environment and maximizing the levels of food production. Thus, the use of sensors in the control and prediction models of agricultural and agroindustrial production can significantly reduce the consumption of agrochemicals and the water used in agricultural production (Bogue, 2017). In addition, it is important to remember that there are different types of sensors that allow the measurement of different variables of the agricultural field, various meteorological conditions, and the same agricultural machinery. Therefore, the precision of agricultural production systems are considered in the current literature by different researchers, academics, and professionals in the field of agriculture and the sciences of the company not only as one of the best management and production practices but also as the simplest, efficient, and effective way to produce more and better food with lesser resources (Bogue, 2017). In addition, according to the most recent survey published by BIS Research, the global market for GPS devices, sensors, cameras, and other electronic devices used in the control and prediction models of agricultural production, as well as systems associated with the management of agricultural production and the various support services for agricultural and

System Architecture

59

agroindustrial producers, will be around 7.6 billion US dollars for the year 2022, which is equivalent to an annual growth of around 12.7% during the period from 2016 to 2022 (Bogue, 2017). Therefore, this significant increase in economic and financial resources for the acquisition of sensors and other electronic devices by the different companies, for their incorporation into the various models of control and prediction of agricultural production, is a clear signal that indicates a demand for the technical upgrade of agricultural and agroindustrial production and justifies the concerns that exists among the people involved in this field. A sensor is a device that converts a physical phenomenon into an electrical signal. A sensor may be classified as digital and analog. The output of a digital sensor varies between one and zero, which translates into a sensor voltage range. The output for an analog sensor may have any value between its voltage ranges. Its voltage output changes according to the reading from the sensor. Analog sensors are connected to the corresponding input pins on the Arduino Mega 2560 and then converted to digital form. Some sensors have analog-todigital converter embedded to the sensor so the data output is in the form of digital data (Corke et al., 2018). Sensors that do not have an on-board analog-to-digital converter have its data sent in analog form to Arduino, which then uses its onboard converter to convert data to digital. After the data are processed to digital form, further processing is on the microcontroller. The sensor or transducer should be compatible with its intended applications. The sensor data are collected, processed, and then transmitted by the sensor node to the receiver by using XBee PRO Series 1 radios. Cost, accuracy, resolution, compatibility with the microcontroller, and ease of use are the criteria on which the sensor is chosen for this chapter’s model.

60

Intelligent Agriculture

2.2.1.1. Solar Radiation Sensor SP-212 Pyranometers are sensors that measure the global shortwave radiation. The SP-212 sensor, specifically calibrated for the detection of solar radiation, provides at its output a voltage proportional to the intensity of the light in the visible range of the spectrum. SP series pyranometers are silicon-cell pyranometers, which are sensitive only to a portion of the solar spectrum, approximately 3501100 nm (approximately 80% of the total shortwave radiation is within this range). Silicon-cell pyranometers are calibrated to estimate the total shortwave radiation across the entire solar spectrum. The SP212 is an amplified, analog sensor with a 0 V to 2.5 V output. Typical applications include shortwave radiation measurement in agricultural, ecological, and hydrological weather networks. 2.2.1.2. Humidity and Temperature Probe Sensor HMP60 This is an analog sensor that provides a voltage output proportional to the relative humidity and temperature in the atmosphere. HMP60 humidity and temperature sensor is a simple and cost-effective sensor for remote data logging applications. This sensor measures air temperature in the range 40°C to 60°C, and relative humidity in the range 0 to 100% RH. Its chip is field replaceable, which eliminates the downtime typically required for the recalibration process. 2.2.1.3. Soil Temperature and Moisture Sensor 5TM This sensor delivers temperature information, measured by an on-board thermistor, along with data on accurate volumetric water content (VWC). This sensor communicates using two different methods, Serial (TTL) and SDI-12 and determines VWC by measuring the dielectric constant of the soil using capacitance or frequency domain technology.

System Architecture

61

Signal filtering minimizes salinity and textural effects, making the sensor accurate in most soils and soilless media. 2.2.1.4. Electrical Conductivity Sensor GS3 The GS3 measures water content, temperature, and electrical conductivity (EC) independently. Its 70-MHz frequency minimizes salinity and textural effects, making it accurate in most soil. Temperature is measured with an on-board thermistor, and EC is measured using a stainless steel electrode array. This is a digital sensor which communicates three measurements over a serial interface and supports SDI-12 serial communication protocol. 2.2.1.5. Soil Moisture Sensor EC-05 The EC-5 soil moisture sensor determines VWC by measuring the dielectric constant of the media using capacitance or frequency domain technology. Its 70-MHz frequency minimizes salinity and textural effects, making this sensor accurate in almost any soil. It gives a wider range of EC measurement and an increased temperature range. The steel needles not only improve sensor contact but they also improve the sensor’s ability to measure EC in porous substrates such as peat. Even though there are a variety of sensors available in the national and international markets that allow the measurement of many meteorological variables and agricultural crops, the use of these five sensors was considered pertinent to incorporate them into the control and prediction model of agricultural and agroindustrial production, considering the information and the available data on the part of the agricultural producers and agroindustrial companies. In addition, after the development of various mathematical analyses with the available information and incorporating in them different variables, both meteorological and

62

Intelligent Agriculture

regarding broccoli culture itself, the results obtained indicate that these five sensors are essential to obtain the best level of production performance. Moreover, it is important to establish that with the use of these five sensors in the control and prediction model of agricultural and agroindustrial production, the agricultural and agroindustrial producers of the state of Aguascalientes (Mexico) would have a higher productive performance than they currently have, in addition to significantly reducing the consumption of drinking water and the various agricultural supplements required for the production of broccoli. Tables 4 and 5 show the technical specification of each sensor used in this project. To provide a greater understanding of the sensors that will be used in this project, it is pertinent to explain the division of the sensors into two large groups: atmospheric sensors and soil sensors. Table 4 presents in detail the technical specifications of the atmospheric sensors that will be used in the control and prediction model of agricultural and agroindustrial production, including even the cost of the same as is available in the webpages of some companies. This would help the reader have a presumption regarding the expenditure an agricultural or agroindustrial product would incur. This knowledge can help take advantage of the existing technology in the market and incorporate it into their productive processes, thereby improving the level of performance and the production of food. With respect to the sensors that measure some of the conditions of agricultural land, Table 5 shows the technical specifications of these types of sensors, also including the price so that readers have an idea of the total costs that they would incur for their agricultural or agroindustrial products. Also, it is important to clarify that the listed sensors do not cover all the existing sensors that measure the conditions of the

Solar Radiation

SP-212: Amplified 02.5 V Pyranometer Apogee PYR Total

Sensors

Solar Radiation Power Supply Sensitivity

Self-powered 2

0.2 mV/µmol m s

Vaisala hmp60 Humidity and

5tm Soil Mositure and Temperature

Temperature Power Supply 5-VDC

3.6 V levels or SDI-12

Measurement Frequency

(water) EC: 023 dS/m

0%100%

System Architecture

Table 4. Technical Specification of Atmospheric Sensors.

(bulk) Temperature: 40°C60°C Spectral Range

410 nm to 655 nm

Temperature Range 0.06% ± 0.06%/°C Accuracy

±5%

Resolution

NA

a: 0.1 εa (unitless) from 20 to 80

Sensor Type

Dielectric

FDR

Accuracy

±7% RH

Electrical conductivity: ± 10% from 0 to 7 dS/m, user calibration required above 7 dS/m

Cost

£260

Cost

£173

£80

63

64

Table 5. Technical Specification of Soil Sensors. Components

Accuracy

Measurement

Resolution

Frequency

Sensor

Power

Type

Supply

GS3 soil mositure,

Electrical conductivity: ±5%

A: 1(air) to 80

VWC: 0.002 m3/m3 (02%

temperature, and EC

from 0 to 5 dS/m, ±10%

(water)

VWC) from 0% to 50%

3.6 V level

Sensor

from 523 dS/m

EC: 025 dS/m

VWC, 0.001 m3/m3 (0.1%

or SDI-12

(bulk)

VWC) > 40% VWC

FDR

Serial TTL,

Cost

£250

Temperature: 40ºC to 60ºC Decagon EC-05 soil

Mineral soil: ±3% VWC

0%100%

NA

Dielectric

moisture

53.6 VDC

£80

@10Ma Electrical conductivity:

εa: 1(air) to 80

a: 0.1 εa (unitless) from 1

temperature

±10% from 0 to 7 dS/m;

(water)

to 20, < 0.75 εa (unitless)

user calibration required

EC: 023 dS/m

from 20 to 80

above 7

(bulk)

FDR

3.6 V level

£80

or SDI-12

Temperature: 40ºC60ºC HMP60 Vaisala humidity And temperature

±7% RH

0%100% RH

NA

FDR

528 V

£173

Intelligent Agriculture

5TM soil moisture,

System Architecture

65

ground; rather, they are those sensors that are considered pertinent to incorporate the control and prediction model of agricultural and agroindustrial production, according to the results obtained from the application of the different mathematical models.

2.2.2. Sink Nodes The receiver nodes play a fundamental role in the connectivity between the sensors and the WSN, which generally manages the information generated by the sensors and establishes a communication with the central or terminal node to send the collected information (Tian et al., 2013). Likewise, the receiving nodes are responsible for sending the commands from the terminal of the data (such as the query of the information or data and the distribution of the ID addresses that are authorized to obtain the information) and receiving the data and information directly from the sensors. Therefore, receiving nodes are fully designed to process the information quickly and expeditiously, process large amounts of data or information, send data or information from great distances, consume a low level of energy, and, commonly, have a lowcost acquisition (Tian et al., 2013). The general structure of a receiving node is presented in Figure 3, which is practically divided into four parts: the microcontroller unit (MCU), the power or energy module required by the node, the communication module, and the module of storage and visualization of information. These four elements are more than enough to understand and analyze the structure of a basic receiver node; when more information is required and, where appropriate, faster information transmission and communication, then the structure of the receiving node will change and will adapt to the

Intelligent Agriculture

66

Figure 3. Structure of the Sink Node. Power module

Storage and Display

Communication module CC2530 module

SD Card

LCD

LPC 2148 MCU

GSM module Ethernet module

Source: Tian et al. (2013).

speed and storage requirements of the information required by users or researchers. In addition, the receiver node commonly uses both an ARM7 LPC2148 card and a NXP 2005 high-speed microcontroller, which generally operates with a general packet radio service (GPRS) module that seamlessly links to the movable base station, as well as communicates via Ethernet with the terminal to send the data and the information received directly from the sensors (Tian et al., 2013). In addition, the power module provides the power required by the receiver node to operate properly, and communication is made through three essential modules: the CC2530 module, the GPRS module, and the Ethernet module. Finally, the information storage and visualization module is responsible for the storage of all the information and data that are generated, as well as the visualization of all the information in the different formats, according to the needs and requirements of the users or researchers (Tian et al., 2013).

System Architecture

67

Additionally, for this project of control and prediction of agricultural and agroindustrial production, it was considered pertinent to use two nodes that receive the data and the information that is generated, both from the same broccoli plantation and having similar meteorological variables. The structure that the two receiver nodes will have is shown in Figure 4. It shows a block diagram describing a typical sensor node developed in this project. The field signals were collected, processed, and then transmitted by the sensor node, that is, from Arduino Mega 2560 to receiver Raspberry Pi by using XBee PRO Series 1 radios. This sensor node could work as a router. This project also considers sensors for monitoring environmental parameters such as air temperature, relative humidity, solar radiation, and precipitation.

2.2.3. Solar Panels Electric power consumption is one of the most recurrent complications in the WSN. This is because, an external environment that is exposed to the sun, rain, dust, etc. proves problematic for the WSN. As a result, the consumption of electric power by WSN is always a matter of concern among researchers, academics, and professionals in the field of agricultural and business sciences (Tian et al., 2013). Therefore, all the components of the WSN that integrate it with each one of the nodes generally are in rest when they are not gathering or transmitting information; however, they require an electrical supply when they do. Thus, all sensors and receiver nodes should have special components such as, for example, microcontrollers, memory cards, sensors, and radio, among other components, which will generally define the time lapses when it in rest, so that electricity consumption is low.

Intelligent Agriculture

68

Figure 4. Block Diagram of a Sink Node. Humidity Sensor

Temperature Sensor Power Supply 5V/2A

Raspberry Pi 3 Model B

Monitor PC GUI

XBee Pro Series 1 Data Transferred Wirelessly

Soil Moisture Temperature Soil Temperature

XBee Pro Series 1

ARDUINO Mega 2560

Precipitation Humidity Solar Radiation

Electrical Conductivity Power Supply

In the literature, the resting state of the components of the WSN is defined as Sn (k = 0,1, ..., k), where n represents the degree of state of rest that the components will have, S0 represents the activity time, S1 ≈ Sk  1 represents the gradual time of rest, and Sk represents the longest resting time of the components which are turned on each other (Rakhmatov et al., 2003). Also, assuming that an event is detected by a sensor node at any time of a given time Nk, the sensor node Nk will conclude its the information gathering process in a determined time t1, and the next event will occur in time t2 = t1 + ti. Then, Nk will make the decision to make a change in the rest period to Sk of the various activities that S0 has to

System Architecture

69

perform. Thus, each Sk activity will have a level of power consumption Pk and a fixed time to transmit the information collected and return again to a state of resting data by τd, j, τu, and k, whereby the definition of the state of rest of the components established in each of the two nodes will be I > j, Pj > Pi, τd, i > τu, j. In addition, to design of the electric power model necessary for the optimal functioning of all the components of the WSN, using an analytical model of high-energy level was considered pertinent (Rakhmatov et al., 2003). Therefore, the battery that will be used will be rechargeable. It will also provide maximum and minimum voltages as required, so the available capacity of the battery will be the different between the maximum and minimum voltage available. Therefore, the available capacity of the battery will depend on two phenomena: speed capacity effect (a higher discharge rate leads to a lower available capacity) and the recovery effect (the battery voltage recovers while resting; Lahiri et al., 2002). This allows an analysis of the relationship among the current load, the download time, and its corresponding minimum load in a certain time of discharge (Tian et al., 2013). Additionally, for the specific case of this project, as the sensors will be located in the field where there is generally no access to the power grid, the project uses solar panels to reduce power consumption. Moreover, the Arduino Mega 2560 will be put in sleep mode when not in use, so that it consume 5mA or less power (Voltaicsystems, 2018a). This project model uses 9 Watt solar panel with 12,000 mAh power battery (Voltaicsystems, 2018b). Another factor considered was the power consumption of the batteries. The battery life is calculated based on the current rating in milliAmpere (mAh) per hour. Ampere is an electrical unit used to measure the current flow toward the

Intelligent Agriculture

70

load. Battery life or capacity can be calculated from the input current rating of the battery and the load current of the circuit. Battery life will be high when the load current is less and vice-versa (Ncalculators, 2018). The capacity of battery can be mathematically derived from the following formula (Ncalculators, 2018):

Battery Life ¼

Battery Capacity in Milli amps per hour Load Current in Milli Amps per hour

Table 6 presents detailed information regarding the calculations performed to determine battery consumption, the types of cards that will be used in the WSN, and the technical specifications of the solar panels that will be required to provide all the electrical energy that is necessary for the optimal functioning of the WSN and all electronic components. In addition, it is important to note that both the WSN and most of the electronic components and microprocessors that make up the control and prediction project of agricultural and agroindustrial production will be most of the time in the agricultural field collecting the necessary data and information, both to facilitate decision-making and to provide the information required by the prediction model to operate properly. Table 7 shows an estimated calculation of power consumption for each sensor when connected to Arduino Mega 2560 and also compares the voltaic system power bank with other batteries. Voltaic system battery power bank will stay active up to 15 to 30 days. Table 8 shows the approximate power that is expected to be generated when there is good sunshine for over five hours in a day (Voltaicsystem, 2018b). As a rule of thumb, the power generation is estimated by multiplying Watts times

System Architecture

Table 6. Battery Power Consumption Calculation. Components

Consumption A

XBee PRO-Tx

0.05

XBee PRO-Rx

0.25

Solar sensor

0.05

Time or Hour

5

Current Consumption

Total Consumption

or Hour

(A × Hour)

0.083333333

0.004166667 0

4

0.066666667

0.003333333

Solar moisture

0.013

1

0.016666667

0.000216667

Temperature and humidity

0.006

1

0.016666667

0.0001

Soil temperature

0.003

1

0.016666667

0.00005

Electrical conductivity

0.025

1

0.016666667

0.000416667

71

Intelligent Agriculture

72

Table 7. Arduino Runtime from V44 Batteries. Current (mA)

Arduino 

Power

Power

Days

Consumption Consumption per Runtime (Watts)

Day (Watt Hours)

V44

5

0.025

0.6

24.4

25

0.125

3

9.8

36

1.1

Sleep Arduino  Normal Arduino 

300

1.5

High power

Table 8. Technical Specification of Solar Panel and V44 USB Battery. Solar Panel Output

V44 USB Battery

Open circuit voltage

7V

Peak voltage

6V

Peak current

1.5 A

Peak current

9 Watts

Capacity

12,000 mAh, 44 Watt Hours

Output

5 V/2 A, 5 V/1 A dual USB

Input

56 V, 2 A

Battery type

Li-polymer

hours and divided by a loss factor of 2 (W × h/2) (Aqeel-urRehman et al., 2018). Technical specifications for the solar panel output and battery are shown in Table 8. Even though existing literature reveals serious discussions among the researchers, academics, and professionals in the engineering and computer sciences regarding the importance

System Architecture

73

of electric power consumption in WSNs and their respective components, this project has tried to establish the type of consumption in the entire equipment as well as the technical specifications of the battery that would be required, remembering that electric power would be obtained through solar panels that are contemplated to be acquired as part of the same project.

2.3. HARDWARE BRAIN The system proposed in this project utilizes sensor boards which consist of a set of sensors connected to a microcontroller, along with a XBee/ZigBee transceiver, which transmits data that are then retrieved at the base station. This microcontroller processes data from the sensors and transmits the sensor’s output to the base station through data nodes. Table 9 shows the hardware components used in this project model. In order for the reader to obtain a detailed knowledge and understanding of the components that the hardware will consists of, in the following sections an attempt will be made to describe a summary of each component possible detailed and minimum necessary for its understanding, according to the

Table 9. Hardware Components. Raspberry Pi 3 Model B

Base station (microprocessor)

Cost £30

Arduino Mega 2560

Microcontroller (reads values/

Cost $35

data from the sensors) XBee PRO Series 1

Wireless communication radios

Cost £75

to transmit and receive the data Power Supply

9 Watt solar panel with 12,000 mAh battery

Cost $160

Intelligent Agriculture

74

project of the control and prediction of agricultural and agroindustrial production.

2.3.1. Raspberry Pi Raspberry Pi is configured as a credit-sized microcomputer based on Raspbian Linux operating system, which offers less complexity and more affordable solutions for wireless monitoring (Powers, 2012; RaspberryPi, 2018). It is the main element in the field of Internet of Things (IoT). It provides access to the Internet and hence connecting an automation system with a controlling device in a remote location becomes possible. Raspberry Pi is available in various versions. Here, Raspberry Pi 3 Model B is used. Some reasons for choosing Raspberry Pi 3 Model B are as follows: • To customize a design and to equip it with the necessary interface and connectivity features can consume a lot of time. Raspberry Pi comes in a complete package which meets the requirements for this project model. • It has the ability to add simple user interfaces, and it is easy to program the Raspberry Pi to perform dedicated tasks. • It is relatively efficient, low cost, has consistent board format, 10 times faster processing, better connectivity, consumes less power, and is suitable for use in a harsh environment. Raspberry Pi acts as a base station in this system. The sensor nodes that have digital output are directly connected to the board leaving them to provide the necessary data without any delay (Powers, 2012; RaspberryPi, 2018). The temperature senor and humidity sensors are directly connected to the Raspberry Pi. The remaining sensor nodes are interfaced with

System Architecture

75

Table 10. Raspberry Pi3 Model B Technical Specification. Processor

Broadcom BCM2387 chipset; 1.2 GHz Quad-Core ARM Cortex-A53 802;11 b/g/n Wireless LAN and Bluetooth 4.1 (Bluetooth Classic and LE)

GPU

Dual Core VideoCore IV® Multimedia CoProcessor; provides Open GL ES 2.0; hardwareaccelerated Open VG; and 1080p30 H.264 highprofile decode. Capable of 1Gpixel/s, 1.5Gtexel/s, or 24GFLOPs with texture filtering and DMA infrastructure

Memory

1 GB

Operating System

Boots from micro-SD card, running a version of the

Power

Micro-USB socket 5V1, 2.5 A

Ethernet

10/100 BaseT Ethernet socket

Linux operating system

Video Output

HDMI

Memory Card Slot

Push/pull micro-SDIO

GB, gigabyte; HDMI, high-definition multimedia interface; SD, secure digital; SDIO, secure digital input output; USB, universal serial bus.

the Raspberry Pi through Arduino Mega 2560. Table 10 shows the technical specifications of Raspberry Pi 3 Model B. As can be seen in Table 10, the technical specifications of the Raspberry Pi3 Model B microprocessor are totally basic, like any processor with these characteristics, and can be purchased basically anywhere in the world because there are several companies that offer them, as much in the market internal as in the international market. Therefore, it is important that the technical specifications of the Raspberry Pi3 microprocessor are presented as thoroughly as possible so that there is a broader knowledge of the type of components

76

Intelligent Agriculture

that will integrate the hardware of the project for the control and prediction of agricultural and agroindustrial production. 2.3.2. Arduino Mega 2560 This project proposal focuses on developing a WSN system by using an open-source hardware platform. Arduino Mega 2560 was found to be well suited for this project. Some reasons for choosing an Arduino Mega 2560 are as follows: • Designing a system from the beginning could consume a lot of time but the biggest advantage of Arduino is that it is in a ready-to-use structure. Arduino comes in a complete package, which includes the 5 V regulator, an oscillator, a microcontroller, serial communication interface, lightemitting diode (LED), and headers for the connections (Jaber, 2017). • Arduino offers several microcontroller models with different characteristics. Moreover, Arduino Mega 2560 boards have more input/output (I/O) pins. These additional pins allow additional hardware devices to be connected (Jaber, 2017). • It is flexible, offers a variety of digital and analog inputs, serial peripheral interface (SPI) and serial interface, and digital and pulse width modulation (PWM) outputs (Jaber, 2017). • It is easy to use, connects to the computer via USB and communicates using standard serial protocol, runs in standalone mode, and as an interface connected to computers. • Arduino controllers are relatively efficient, consume less power, are low cost, and suitable to use in harsh environments (Jaber, 2017).

System Architecture

77

Arduino is an open-source tool for developing computers that can sense and control more of the physical world than a desktop computer. It is an open-source physical computing platform based on a simple microcontroller board, and a development environment for writing software for the board. The software is written in C or C + + programming language (Arduinomega, 2018). The Arduino development board is an implementation of wiring, a similar physical computing platform, which is based on the processing multimedia programming environment (Methley et al., 2008; Sheikh & Xinrong, 2018). It is a microcontroller board based on the ATmega1280. In WSN, it acts as a medium to the sensor for providing data to the Raspberry Pi which acts as a base station. The sensors having analog output are usually interfaced with the Arduino analog pins (Methley et al., 2008; Arduinomega, 2018). The output from the sensor node can also be compared with the voltage that can be given to analog ref pin. Arduino XBee Shield helps Arduino to communicate via radio frequency (RF) transceiver (Sheikh & Xinrong, 2018). It mounts directly on Arduino and holds XBee RF transceiver. The data gathered by the sensors nodes are passed on to the base station. Table 11 shows the technical specification of Arduino Mega 2560.

2.3.3. XBee PRO Series 1 In this system, Digi International XBee PRO Series 1 is used. XBee PRO Series 1 60mW is a wireless module that is good for point-to-point and multipoint, and is convertible to a mesh network point. XBee PRO is a proprietary protocol stack that provides a network layer and the framework for an application layer over the physical (PHY) and media access control (MAC) layers of IEEE 802.15.4 (ZigBee,

Intelligent Agriculture

78

Table 11. Arduino Mega 2560 Technical Specification. Microcontroller

ATmega2560

Operating Voltage

5V

Input Voltage (recommended)

712 V

Input Voltage (limit)

620 V

Digital I/O Pins

54 (of which 15 provide PWM output)

Analog Input Pins

16

DC Current per I/O Pin

20 mA

DC Current for 3.3V Pin

50 mA

Flash Memory

256 KB of which 8 KB used by boot loader

SRAM

8 KB

EEPROM

4 KB

Clock Speed

16 MHz

2006). Also, communication between the sensor board and the base station is done through XBee transceivers. These are much more powerful than the plain XBee modules, which is great when one needs more range. XBee provides low rate, low-power, and low-cost wireless networking. IEEE 802.15.4 defines the XBee standard using MAC protocol (RaspberryPi, 2018). IEEE 802.15.4 is a technical standard which defines the operation of networks. The outdoor communication range of XBee PRO Series 1 is around one mile, and it operates at 24 GHz (RaspberryPi, 2018; ZigBee, 2006). The main use of IEEE 802.15.4 (MAC and PHY layers) is for communication with low-power and lowthroughput requirements (RaspberryPi, 2018; ZigBee, 2006). The objective of this protocol is to have reliability when transferring data, simple installation process, reasonable battery life, low cost, and a flexible protocol.

System Architecture

79

Figure 5. IEEE 802.15.4 Stack.

Upper Layers MAC PHY

Figure 5 illustrates the protocol stack of the IEEE 802.15.4 standard. This protocol stack is based on the open systems interconnection (OSI) seven-layer model, and it corresponds to the bottom two layers, that is, the PHY and the MAC layers (Jin-Shyan et al., 2018; ZigBee, 2006). In this system, the XBee configured with the base station will act as coordinator whereas another one on the receiving side will act as router. The communication with the base station and computer is done through UART communication. It provides both AT and application programming interface (API) mode serial interfaces for communication. In AT mode, XBee module performs as a serial replacement. When operating in API mode, the data entering and leaving the module is divided into frames that define operation (Jin-Shyan et al., 2018; ZigBee, 2006). Additionally, ZigBee is one of the most used applications of IEEE 802.15.4 standard, because compared to other

80

Intelligent Agriculture

communication protocols of small ranges, such as Bluetooth or Wi-Fi, ZigBee is the most suitable application for the common areas that are remote in external environments, particularly for monitoring applications of various meteorological variables and agricultural crops (Tian et al., 2013). Therefore, the ZigBee protocol is the most widely used because it has been developed to consume less electrical power and provide wireless connectivity for a variety of applications and microprocessors. Various researchers, academics, and professionals in the field of computational sciences consider that the ZigBee is the standard protocol that best fits the communication link between the various components of the hardware. In this sense, in the network in which ZigBee generally operates, there are three types of nodes: the coordinator (receiving node), the router, and the final device. All ZigBee networks can only have one coordinator node, which can select the frequency channel, the initial networks, and all other nodes with which it can communicate, as well as with the other devices. Both end devices as the router are commonly used to send or receive data or information collected by the sensors (Tian et al., 2013). In a different way, the router can also read and send messages to all the subnodes or sensors that are connected in the network, which facilitates a higher percentage of communication between all the components that make up the hardware, thereby transferring the data and information in a faster and more efficient manner.

CHAPTER 3 SOFTWARE AND APPLICATIONS

3.1. INTRODUCTION The development and use of software in the prediction system of industrial and agricultural production has increased significantly in the last two decades and are a fundamental part of the agroindustrial technological development. One of the main reasons for this significant growth may be the high demand that companies have on the accuracy and functionality of the expected results in the production prediction models. Therefore, the new generation of industrial robots and industrial technology offer companies much more accurate control of production prediction levels than any other traditional way, which allows agribusiness companies to improve not only the business results obtained but also the acquisition of a higher level of growth and business performance. In this sense, the software incorporated into the agroindustrial technology has high levels of precision for production prediction because they incorporate a series of very complex algorithms, which facilitate the presentation of results in relatively shorter periods with greater relevance. 81

82

Intelligent Agriculture

Therefore, currently, the software incorporated into the agrotechnology generally offers greater functional advancement than the equipment that was used a few years ago. Thus, a software offers to agroindustrial companies, primarily, a greater expectation of obtaining more and better business results because it usually has some flexibility, depending on certain environmental conditions, and is capable of generating various scenarios for predicting agroindustrial production and the presentation of the results in different ways. In addition, agrotechnology with incorporated software is increasingly becoming a powerful tool in agroindustrial production. This is because they help companies to be more efficient and effective in the acquisition and generation of data, as well as in the analysis and prediction of production through complex mathematical models, which helps solve the problems arising because of overproduction in agroindustries. Therefore, software and agrotechnology commonly work in unstructured and unpredictable environments, which allows them to react in a relatively short time and intelligently to external factors such as wind, solar radiation, humidity, even in extreme conditions, which become practically fundamental factors for the prediction of agricultural and industrial production. Under this scenario, the development of agricultural production typically requires the use of software for the optimal functioning of agrotechnology, as well as for the different purposes of prediction in agroindustrial production. This is because the softwares obtain diverse information through the sensors that are incorporated into computers for the acquisition or collection of data, the processing of said data, and the communication of the resulting information by means of different graphs and reports. Therefore, any software incorporated into agrotechnology

Software and Applications

83

generally contains a low level of control; it accepts a series of values that coordinate to present information that predicts agricultural production, which allows not only the reduction of errors while planning agroindustrial production but also taking advantage of the overproduction that could be generated in agricultural fields. Thus, program control of the software is typically done using a series of drivers that communicate directly with the sensors through a series of algorithms that are too sophisticated to facilitate the activities of planning, control, analysis, and presentation of the information collected. Therefore, an advanced system with agrotechnology and software incorporated in it could offer agroindustrial companies a series of programs for the configuration of production systems, calibration, and analysis of the data generated in the system. In addition, configuring the production systems will make it easier for agribusiness companies not only to improve their agricultural and industrial planning system but also to significantly reduce the waste of food that could be generated by agricultural overproduction. In this sense, software and agrotechnology are becoming an essential part in the growth and development of companies, mainly in the agroindustrial industry, because it facilitates the generation of possible scenarios in the prediction of agricultural and agroindustrial production. This is accomplished through the various modules and sensors that are incorporated into the system, which obtain valuable information from a series of factors or variables essential for the agricultural production process. Therefore, if agribusinesses want to significantly improve their production prediction systems, one of the best alternatives is the agrotechnology with incorporated software that is being developed nowadays.

Intelligent Agriculture

84

3.2. SOFTWARE Farmers need a predictive software application in order to decrease the error range that might arise during crop processing. Besides, its use will enable efficient management practices and the generation of information that will guarantee the sustainability of resources and the maximization of harvest results. This application helps to reduce the margin of error within the harvest prediction process, on the one hand, and allows the user to make appropriate decisions with the exploitation of the information contained in the database (DB), on the other. This results in the harvest process becoming efficient, and likewise, the optimization of the resources. Some aspects of the software application in its final version are shown through screenshots. Registration, sensor administration, monitoring of field variables, and forecasting are some of the modules this application will have as shown on the forthcoming figures. Table 12 shows the progress made up to date in terms of the Smart Farming software application. Table 12 contains an actor’s description and functions for the system. These actors are responsible for interacting with the system: the administrator and the user manipulate the system, wireless sensor network (WSN) connects to the system by sending a file with variables measured by the sensors, and the weather station interacts indirectly through a file with atmospheric variables. User cases represent a functional view of the system being developed. These portray both, a general overview, and an informational detailed use. The former is shown through simple user case diagrams, Figures 6 to 16, which illustrate the main functionality of a system and the actor that interacts with it. The latter is shown through informational tables, Tables 12 to 17, expanding on what is shown in the user

Software and Applications

85

Table 12. Actor Description/Functions. Actor Normal user

Description/Functions Any type of user whose purpose is the crop forecast. Preferably, a farmer or person is in charge of the plot Activities: • Historic data graphs generation (DB) to visualize forecast • Visualize important information about the crop (DB variables)

Administrator user

Any type of user who performs the administration of the application for the forecast of the harvest Activities: • Historic data graphs to visualize forecast, as well as calculate linear regression • Display relevant information of the harvest (variables of the DB) • Manage the sensors connected to the network of nodes, as well as view the information relative to them • Manage users of the application: registrations, cancellations, and queries • Fill in the DB with the data of the variables of interest from the file generated by the sensor network

WSN

Sensor network that captures the values of some variables (through sensors) to generate an excel file

Weather station

Measures some environmental variables that are considered for the forecast of the harvest that are saved in an Excel file

Intelligent Agriculture

86

Figure 6. Business Use Case. Prediction

Crop

Filling DB

Weather Station

User User login Manage sensors WSN Administrator

Employee administration

case diagrams by incorporating data related to the identification of each stakeholder, its corresponding description, and its function within the system where applicable. Figure 6 portrays the business user case. The graph details the main functions of the system represented by the ovals, as well as the direct (manipulate the system) and indirect actors of the system (they send information useful to the system) represented by the people’s charts. Lines between functions and actors represent the interaction between them. Figure 7 explains the prediction crop function performed by the software, and it can be observed that the prediction of food production (broccoli) is represented through various graphs and tables in the results or output of the information. Table 13 clearly describes the process carried out by the software for predicting food production (broccoli), which is fed by the information that is collected in the different time areas of agricultural production. Likewise, Figure 8 shows the historic data process, highlighting the gathering of the historical information of the

Software and Applications

87

Figure 7. Prediction Crop Use Case. User login

Prediction DB query Administrator User Results display

Figure 8. Display Historic Data Use Case. User login

Crop

DB consult

DB query

Administrator

User Screen display

agricultural production and agroindustry from the DB. The more robust the DB of the historical agricultural production of agroindustry is, the better results it will yield. In addition, Table 14 accurately describes the process of processing historical information on agricultural production, as well as the presentation of it in various formats

Intelligent Agriculture

88

Figure 9. Manage Sensor Use Case.

Visualize sensor

User login

Administrator Consulting BD

Figure 10. Filling Database Use Case.

User login

Administrator Weather Station

Open data raw file

Read raw file WSN

Filling DB

according to the needs of users. Therefore, as long as the historical information on agricultural production is greater, the prediction of production and its results will be much more robust.

Software and Applications

89

Figure 11. Employee Administration Use Case.

User login

Search

Administrator Add employee

Delete employee

Edit employee

Figure 12. Prediction Menu Sequence. User

Administrator System

Prediction menu

Type user/login () No user () Type user/login () No user () Menu variable and graphic type () Menu variable and graphic type () Display graphic () Display graphic ()

Figure 9 illustrates how to manage a sensor user case, which will help visualize the sensors that make up the system. Its constant monitoring makes it possible to predict any type of problem they may arise. By keeping the operation at

Intelligent Agriculture

90

Figure 13. Crop Menu Sequence. System

Crop menu

User Type user/login () No user () Type user/login () No user () Menu variable and graphic type () Menu variable and graphic type () Display information () Display information ()

optimal, the prediction of agroindustrial production is possible without any problem. Table 15 describes the management processes of the sensors that make up the prediction system of agroindustrial production, which is essential to constantly monitor the proper functioning of the sensors and ensure that they are providing the information required for the prediction of agroindustrial production. Figure 10 presents the filling DB processes, by means of which the information provided by the meteorological station is collected, whether it is obtained through a station in the same agribusiness agricultural field or through an Internet connection with the meteorological station of the locality or community where the agroindustrial company is located, in addition to the information provided by the various sensors that make up the production prediction system.

Software and Applications

91

Figure 14. Manage Sensor Sequence. Administrator System

DB

User login ()

User error ()

Query ()

Display information ()

Figure 15. Filling Database Sequence. Administrator System

Raw file

DB

User login () User error () Open file () Query () Filling DB () Success mess ()

Intelligent Agriculture

92

Figure 16. Employee Management Sequence. Administrator System

Employee menu

DB

User login () User error ()

Choose option () Query () Result action () Display information ()

Action mess ()

Table 16 describes the process by which the information provided by the meteorological station installed in the agricultural fields of the agroindustry is obtained. This may be through an Internet connection, apart from the information obtained from the various sensors that are installed in the production prediction system. Figure 11 illustrates employee administration, wherein the person responsible for systems or planning of the agroindustry itself will be the one who efficiently manages the software. Moreover, this is the same person who is able to grant all the corresponding permits for other users to have access to information on the prediction system of agroindustrial production. Table 17 describes the process involved in the administration of the software, which will be managed by an employee of the agroindustry and he/she will be able to grant corresponding permissions to other employees or workers of the organization so that they can enter and obtain the information that generates the prediction system of agroindustrial production.

Software and Applications

93

Table 13. Prediction User Case Description. Use case name:

Prediction

ID:

1

Importance

High

level: User Primary Actor:

Use case type:

UC1

Administrator Stakeholders and interests: User: Needs to see historic variables behavior in order to “predict” future crop. Administrator: Needs to see historic variables behavior in order to “predict” future crop. Brief description: Both stakeholders, user and administrator, can consult graph (different types of graphs) the behavior of historical variables harvest in a specific date, in order to forecast crop. Tigger: Relationship: Association: User; Administrator Include: User login Extend: Generalization: Normal Flow of events: 1. User /administrator is identified in the system (login). 2. User / administrator chooses historic variables and the kind of graphic (bar chart, pie chart, scatterplot). 3. System will show historic variable graphic chosen, according to a specific date. Sub flows: Alternate/exceptional flows:

Intelligent Agriculture

94

Table 14. Crop User Case Description. Use case name:

Crop

ID:

2

Importance

High

level: User Use case

Primary Actor: Administrator

type:

Stakeholders and interests: Administrator and user choose temperature (temperatura), humidity (humedad) or rainfall (precipitación) variables stored in the data base, to see their behavior into a graphic. Brief description: Both stakeholders, user and administrator, can consult the behavior of historical variables harvest through a graphic (bar chart, pie chart, scatterplot). Tigger: Relationship: Association: User, administrator Include: User login Extend: Generalization: Normal Flow of events: 1. User /administrator is identified in the system (login). 2. User / administrator chooses temperature (Temperatura), humidity (Humedad) or rainfall (Precipitación) variable to see it´s values in a graph (bar chart, pie chart, scatterplot). 3. System will go to the database to launch the query that allows to exploit the database that contains the information. 4. System will graphic information from variable chosen. Sub flows: Alternate/exceptional flows:

Software and Applications

95

Table 15. Manage Sensor User Case Description. Use case name:

Manage

ID:

3

Importance

sensor Primary Actor:

Administrator

High

level: Use case type:

Stakeholders and interests: Administrator: activates the filling data base system procedure. Brief description: Administrator need to consult sensor information given by the WSN in the raw data file. Tigger: a. Consulting system DB Relationship: Association: Administrator Include: User login, Consulting DB. Extend: Generalization: Normal Flow of events: 1. Administrator logs the system (login). 2. Administrator chooses manager sensor option 3. Variable sensors are shown. Sub flows: 3.1. DB sensor query is launched when administrator chooses Sensor option Alternate/exceptional flows:

Intelligent Agriculture

96

Table 16. Filling DB User Case Description. Use case name:

Filling DB

ID:

4

Importance level:

High

Weather Station Primary Actor:

Use case type: WSN

Stakeholders and interests: Weather Station: sends information about environmental variables temperature, rainfall, humidity, etc. This information is stored in a raw data file. This file is send to the system. WSN: through the sensors connected to it, generates a file (raw file) containing the values detected by each sensor that will feed the system database. Brief description: Weather Station sends raw file with environmental variable values in order to fill the database. Similarly, WSN generates values of variables measured by the sensors connected to it. These information is stored in a raw file. Both raw files, are opened before the administrator and user consult, as well as the filling data base system process. Tigger: 1.1. Filling DB Relationship: Association: Weather Station, WSN Include: Open Data Raw File; Read Raw File Extend: Generalization: Normal Flow of events: 1. Administrator logs the system (login). 2. Administrator fills the system data base with the data raw file, through the “Actualizar información” button. 3. System notifies the administration when filling of the DB systems process has been completed. Sub flows: 1.1. Data base query is performed in order to fill data base system. Alternate/exceptional flows:

Software and Applications

97

Table 17. Employee Administration User Case Description. Use case name:

Employee

ID:

5

Administration

Primary Actor:

Administrator

Importance

High

level:

Use case type:

Stakeholders and interests: Administration: sends information about environmental variables temperature, rainfall, humidity, etc. This information is stored in a raw data file. This file is send to the system. Brief description: Administration sends raw file with environmental variable values in order to fill the database. Similarly, WSN generates values of variables measured by the sensors connected to it. These information is stored in a raw file. Both raw files, are opened before the administrator and user consult, as well as the filling data base system process.

Tigger: Relationship: Association: Administrator Include: User login Extend: Generalization:

Normal Flow of events: 1. Administrator is identified in the system (login). 2. Administrator chooses whether to perform a search, add a new employee, edit an existing employee, or remove it from the database. 3. System show messages according to each action. Sub flows: 3.1. a Search process can be done using name/last name or job title. 3.2. b Depending on the search, information is show on the system interface.

Alternate/exceptional flows:

98

Intelligent Agriculture

3.3. SOFTWARE DESIGN Entity-relationship diagram (ERD) substitutes the class diagram in this chapter. An ERD is a data modeling technique that graphically illustrates an information system’s entities and the relationships between those entities. This diagram is a conceptual and representational model of data used to represent the entity framework infrastructure. The elements of an ERD are as follows: • Entities: It is a thing that exists either physically or logically and can be uniquely identified. The entity is an abstraction from the complexities of a domain and is a definable thing or concept within a system. When one speaks of an entity, one normally speaks of some aspects of the real world that can be distinguished from other aspects of the real world. • Relationships: It captures how entities are related to one another. It also describes the association or interaction between entities. • Attributes: It is a particular property that describes the entity. The attribute is a property or characteristic of the entity that holds it. An attribute has a name that describes the property and a type that describes the kind of attribute it is, such as varchar for a string, and into for integer. ERD of this chapter shows the entities involved in the system, the attributes that define them, as well as the relationships between them. A sequence diagram is a type of interaction diagram. It contains objects that participate in a user case and the messages that pass between them over time. This diagram is a dynamic model that shows explicit sequence of messages that are passed between objects in a defined interaction. It is important to clarify that the sequence diagrams were made

Software and Applications

99

based on the descriptions of the user cases presented in this same section. The sequence diagrams of each user case description are presented in the forthcoming figures. Figure 12 illustrates the prediction menu sequence. The figure clearly presents the sequence that should be followed, both for a user and for the administrator of the system, for the prediction of agroindustrial production. Figure 13 illustrates the crop menu sequence, in which the sequence to be followed in the prediction of broccoli cultivation can be clearly observed. This helps to feed data in the system for the prediction of agroindustrial production. Figure 14 illustrates the manage sensor sequence. It presents the sequence that will be followed to establish software connection with the sensors, which will provide important information for the prediction of agroindustrial production. Figure 15 shows the filling DB sequence. This sequence makes it possible to observe the sequence of the management of the sensors with which the software and the system will work in general, which will provide the necessary information for the prediction of the agroindustrial production. Figure 16 illustrates the employee management sequence. Herein, the sequence to manage the employees is established. These employees will have access to the use and management of the system, as well as those responsible for the feeding of the same information. Human–computer interaction (HCI) studies the way people interact with computers and to what extent computers are or are not developed for successful interaction with human beings. When the concept of HCI is integrated into the interface design, different mental processes and learning that users possess, such as their needs, their desires, their cultural and behavioral aspects, as well as their physical limitations, should be incorporated. Smart monitoring and control system (SMCS) system considers the application of the HCI concept

100

Figure 17. Entity-Relationship (ER) Database.

Intelligent Agriculture

Software and Applications

101

as well as the design principles that support it in its development. Moreover, the application was evaluated with real users (not from the adequate sample) and their corresponding adjustment (based on the results of the evaluation). The ERD is a data model used to analyze and describe the data requirements and assumptions in the system from a top-down perspective. This diagram represents how information will be stored through entities (15 in total) and its attributes (data collected on each table), as well as the structure needed to draw information from multiple entities called relationships. The ERD for the Smart Farming DB is shown in Figure 17.

3.4. APPLICATION SCREENSHOTS The normal user and administrator can see the available options. If the administrator user wants to login, the Aceptar

Figure 18. Administrator/User Access.

102

Intelligent Agriculture

button must be selected. However, if the normal user wants to login No soy un administrador button must be pressed. Figure 18 shows the access to the software application. Three buttons are shown: Inicio, Predicciones, and Cosecha. Two different users can use the application, that is, the normal

Figure 19. Administrator Menu.

Figure 20. Predictive Menu.

Software and Applications

103

user and the administrator. The Acceso Administrador button allows the administrator user to have access (please note that the screenshots in the figures are in Spanish as they are intended for use in Mexico). Figure 18 shows the screen used to access the prediction system of agroindustry production. The system database will hold a record of the employees that the agroindustry company considers necessary to have access to the system and its information. Every person who enters the system will be

Figure 21. Formula Menu Button.

Figure 22. Maximum Rainfall Bar Graph.

104

Intelligent Agriculture

given a username and password, which will allow them to only have access to certain information that the agroindustry considers prudent. Figure 19 shows the administrator menu, displaying the Cosecha, Predicciones, Sensores, Personal, and Actualizar Informacion buttons. These buttons will allow the users to

Figure 23. Rainfall Probability Dispersion Graph for the Next Seven Days.

Figure 24. Rainfall Probability Pie Graph for the Next Seven Days.

Software and Applications

105

obtain information of the past 19 years pertaining to broccoli harvests, including data of the current crop that is being grown in the different agricultural production plots. Moreover, the production level of the current broccoli crop is also predicted according to a series of variables that are measured from time to time. The administrator and the normal user can use the Predicciones button. Its objective is to allow the user to choose

Figure 25. Cosecha (Crop) Menu Option.

Figure 26. Sensor Menu Option.

106

Intelligent Agriculture

the period during which prediction must be performed; this is achieved by incorporating dates (historical data are included) from a calendar as shown in Figure 20. The Principal button available in all system interfaces allows both users to access the principal menu. The Ayuda button helps users interact with the system.

Figure 27. Employee Button Menu.

Figure 28. “Agregar Personal Nuevo” Button.

Software and Applications

107

Figure 21 illustrates how to access the regression model for predicting crop yield; its usage is restricted exclusively for the administrator. This is a preliminary prototype and only the linear regression is presented. Both the administrator and normal user have access to the Predicciones menu, which permits choosing variables such as

Figure 29. “Editar Personal” Button.

Figure 30. “Eliminar Personal” Button.

108

Intelligent Agriculture

Temperatura, Precipitación, and Humedad, as well as the chart type for graphical representation. Figures 2224, show different types of variables that are graphically represented. Both users can also use the Cosecha button. Figure 25 shows relevant information about the current crop such as humidity, amount of rainfall, and temperature. The Actividades para hoy section allows both users to remember activities planned for the day. Figure 26 shows sensor information and iterates that the administrator is the only user. The screen shows a scale with the current status of the sensors that are in the crop, as well as detailed information regarding the last time it received data pertaining to maintenance, temperature, location, etc. Figure 27 illustrates how the administrator can control the employee requirements as Ingresar un nuevo empleado and Dar de baja a un empleado, as well as look for some specific employee information. The interface shows the personal data information, and the possibility to find an employee by searching using name/last name or job title. “Personal” menu shows three buttons for personnel administration. Figure 28 shows the process to add new personnel information (Agregar Personal Nuevo button), Figure 31. “Ayuda” (Help) Button.

Software and Applications

109

Figure 32. Normal User Menu.

Figure 29 shows the process to edit personnel information (Editar Personal button), and Figure 30 shows the process to remove personnel information (Eliminar Personal button). Both the administrator and normal user can see the Ayuda button. Figure 31 shows relevant help information to use the system correctly. Figure 32 shows the normal menu wherein Cosecha and Predicciones buttons are shown.

This page intentionally left blank

CHAPTER 4 AGROTECHNOLOGY SOLUTIONS

4.1. INTRODUCTION Food security and water conservation are arguably two of the biggest challenges currently facing humankind. Therefore, these have emerged as major concerns for many developed and developing nations. In addition, developing countries are facing challenges linked to economic growth and creating sustainable livelihoods (Aqeel-ur-Rehman et al., 2018). In this context, more efficient and environmentally sustainable farming operations can emerge as one of the potential strategies to address these challenges. Thus, addressing inefficiencies, waste, and excessive consumption of natural resources in farming operations will provide direct benefits to nations, for example, by reducing crop and food waste, increasing the efficient use of natural resources (i.e., water, energy, and land), and reducing CO2 emissions and potential pollution of the environment/land created by the excessive usage of fertilizers. These benefits can ultimately assist farmers in improving the productivity and sustainability of their farming operations, resulting in food security and water conservation. 111

112

Intelligent Agriculture

These steps will consequently aid communities by benefiting them socially, economically, and environmentally in both the short and long terms. The Institutional Links initiative will, therefore, bring researchers and industrial experts from both the United Kingdom and Mexico to a common platform to develop a smart monitoring and control system (SMCS) for agrotechnology, which will enable more efficient and sustainable farming operations and combat its associated challenges. The initiative will provide valuable knowledge and lessons for developers, users, and communities in regard to the improvement of farming operations and the development, testing, and implementation of agrotechnologies in developing countries. Therefore, the systematization of these processes will form a key element of the project, thereby building local research capacity and supporting policymakers in the formulation of effective policies and/or strategies to facilitate the development of agrotechnology. The aim of the project is to develop the following components: • a data acquisition (DAQ) system comprising a set of sensor nodes (SNs) and data nodes for monitoring critical variables in plantations (e.g., soil moisture level and temperature, air temperature and humidity, etc.); • a software (SW) application for monitoring and data processing; and • a crop yield-forecasting model. Sensor and data nodes require the design, construction, and commissioning of hardware (HW) and SW. The project will be of interest to small-scale and large-scale farmers encountering uncertain weather phenomena that often lead to loss of produce. This innovative development will be an empowering process for the farmers to resolve

Agrotechnology Solutions

113

issues pertaining to climatic variations. Furthermore, this project may help improve the efficiency of irrigation systems to reduce water usage, reduce electrical energy consumption, reduce CO2 emissions and hence environmental pollution, and also is expected to help in improving crop yield forecasts in order to reduce crop and food wastes, as well as achieve a better supply-to-demand ratio. By monitoring the weather conditions along with other soil variables, the irrigation system is expected to be more efficient because only the required amount of water will be used. Furthermore, the impact of diseases on agricultural production can be controlled or reduced if the health condition of crops is monitored appropriately and the corresponding actions are taken on time. This can also be translated into an expected reduction in the use of pesticides. The use of innovative tools based on digital technologies in farming is expected to bring many benefits, such as increased productivity and profitability. As a result, the associated increase in productivity helps reduce the food security risk faced in some regions of the world, more specifically in Mexico as this is the country considered in this project. Wireless sensor network (WSN) technologies have rapidly developed over the years. WSNs can be used in agriculture to provide farmers with a large amount of information. Precision agriculture is a management strategy that employs information technology to improve quality and production. Utilizing wireless sensor technologies and management tools can lead to a highly effective, green agriculture (Aqeel-urRehman et al., 2018). From several perspectives, field management can improve precision agriculture, including the provision of adequate nutrients for crops and the wastage of pesticides for the effective control of weeds, pests, and diseases. The information available in the public domain outlines the recent applications of WSNs in agriculture research

114

Intelligent Agriculture

as well as classifies and compares various wireless communication protocols, the taxonomy of energy-efficient and energy harvesting techniques for WSNs that can be used in agricultural monitoring systems, and comparison between early research works on agriculture-based WSNs (Aqeel-urRehman et al., 2018; Methley et al., 2008). The challenges and limitations of WSNs in the agricultural domain are explored, and several power-reduction and agricultural management techniques for long-term monitoring are highlighted. These approaches may also increase the number of opportunities for processing Internet of Things (IoT) data (Aqeel-ur-Rehman et al., 2018; Methley et al., 2008). Battery-powered WSNs comprise several sensors, processors, and radio frequency (RF) modules. The SNs or motes can communicate wirelessly through a communication link and forward their data to a base station or coordinator node by communicating with a gateway (Aqeel-ur-Rehman et al., 2018; Corke et al., 2018). The aim of this project is to improve the efficiency and sustainability of farming operations to support food security, water conservation, and hence economic growth in Mexico by developing an SMCS agrotechnology. This project intends to implement a sensor DAQ system comprising a set of sensor data nodes for monitoring critical variables in plantations. Nodes require the design, construction, and commissioning of HW and SW. The project will be useful to small-scale and large-scale farmers to reduce their loss in crop yield while encountering uncertain weather phenomena. This innovative development will empower the farmer to resolve such issues. Moreover, the project will improve the efficiency of the irrigation system, thereby reducing water usage, electrical energy consumption, and CO2 emissions and hence environmental pollution. It will also help in improving the production

Agrotechnology Solutions

115

forecast in order to reduce crop and food wastes, as well as achieving a better supply-to-demand ratio. This project is at technology readiness level 3 (TRL3), which means that it involves a practical design and experimental proof of concept through several analytical studies (e.g., analysis of measurement techniques and their application in farming operations, and data analysis and construction of a mathematical model for yield forecasting) and multiple experimental studies (e.g., WSN testing to validate and verify that all data are acquired and processed as expected, construction and validation of the mathematical forecasting model through computer simulations using field data, comparison of results from simulations and real online data, comparison of results using the current forecasting approach and the proposed model, etc.). It includes analytical studies (e.g., analysis of current farming operations and different sensors and electronic board technologies, analysis of field data to build a mathematical model for yield forecast, analysis of performance of the proposed system for proof of concept, etc.). The project also involves validation and verification of designs at the laboratory level. The University of Derby (UoD) will develop the HW prototypes, as well as support the development of the SW prototypes. This will be enabled by UoD’s expertise on industrial operations improvement and electrical and electronic engineering. Universidad Autonoma de Aguascalientes (UAA) will contribute to tasks mainly related to SW deployment. La Huerta will provide industrial knowledge and expertise along with historical data from their plantations. All TRL3 activities of the project support the proof of concept of the smart monitoring and control system (SMCS) (hardware (HW) and software (SW)). All laboratory-based activities were initially carried out at University of Derby (UoD) and UAA, and afterward in relatively small agricultural

116

Intelligent Agriculture

fields that require HW components/devices for construction of a WSN. The following items are needed to build SNs: development electronic boards (motherboards and daughterboards), sensors (e.g., soil moisture level, air temperature and humidity, etc.), enclosures, and power supplies (batteries, solar panels, etc.). In addition to SNs, data nodes are also needed and they are made up of development boards, a few sensors (e.g., temperature and humidity) to monitor the condition of each unit, enclosures, and power supplies.

4.2. AIM AND OBJECTIVES This project corresponds to a proof-of-concept validation of an electronic sensor board prototype, a WSN prototype including HW and SW, a mathematical model, and SW components for crop yield forecasting to be used within a smart farming framework. The project involves analytical and laboratory studies (e.g., mathematical modeling, analytical derivations, computerbased simulations, and laboratory experiments). It is also expected that physical experimentation may be carried out in relatively small agricultural fields due to the nature of the project. Project key deliverables are as follows: • a smart farming sensor board prototype that will measure and transmit field data (e.g., soil moisture level, soil temperature, leaf wetness, air temperature and humidity, etc.); • WSN prototype comprising a set of sensor and data nodes for monitoring critical variables in plantations. Nodes require the design, construction, and commissioning of HW and SW; • a crop yield-forecasting mathematical model constructed from and validated through historical and online data obtained through the WSN;

Agrotechnology Solutions

117

• construction of SN and WSN prototypes using development boards for sensor and data nodes, sets of sensors for measuring relevant field variables, and wireless communication daughterboards; • deployment of SW components for DAQ and processing of signals from sensors; • statistical analyses of existing historical data from local plantations and determination of past estimation errors between real harvesting quantities and the corresponding estimated values; • evaluation and selection of different model structures (e.g., auto regressive moving average (ARMA)/autoregressive integrated moving average (ARIMA), neural networks, etc.) in order to assess their suitability for the studied problem; • construction of a mathematical model using a subset of field data currently available; • deployment of smart monitoring and control software components (SMCSC) that will allow for supervision and control of field variables and also offer several analysis tools (e.g., water and energy consumption); • feasibility study for analyzing the integration of the yieldforecasting model onto the SMCSC; validation and verification of SW components through testing; • validation and verification of HW and SW prototypes in a laboratory; and • validation and verification of HW and SW prototypes in relevant environments (i.e., local agricultural fields) in the United Kingdom and/or Mexico.

118

Intelligent Agriculture

4.3. PROOF-OF-CONCEPT ACTIVITIES The wireless network architecture consists of the following three elements: a coordinator, router, and end device. A coordinator is responsible for handling and storing data received from the routers. Routers are an intermediary device that permit data to be sent to and from other devices; they request data from an end device and send this data to the coordinator. End devices have limited functionality for communication as they are not responsible for routing data and are used to take measurements upon request; they can remain in sleep mode until a router requests data. These three elements together create a Zigbee network in a cluster tree configuration for the purpose of the project as shown in Figure 33. Multiple Arduino Mega 2560 (acting as Slaves) were used in the smart farming wireless network as an end device and are assigned to separate routers as shown in Figure 40, with approximately three end devices connected to a single router to create a single node. These Arduinos are used to collect data using sensors connected to the general-purpose input/output (GPIO) pins, therefore collating results for electrical conductivity

Figure 33. Smart Farming Zigbee Cluster Tree Network.

Agrotechnology Solutions

119

(EC) of the soil, soil moisture, air temperature, relative humidity, and global radiation. A Raspberry Pi 3 (acting as a Master) is used to operate as a router within a single node and is responsible to make a request for data from the end devices to take different sensor measurements that are allocated to a specific measurement frequency as shown in Table 18. The routers used in each individual node will then receive the comma-separated value (CSV) (see Figures 34 for example) from the end devices containing the measurements

Table 18. Smart Farming Measurement Frequency. Sensors

Site

Soil

Parameter

Electrical

Units

Range

Min

Max

dS/m

0

6

Accuracy

Measurement

(+/)

Frequency

0.2

Daily

conductivity Soil

Moisture

%

0

50

1

Hourly

Air

Temperature

°C

10

40

0.5

Hourly

Air

Relative

%

0

100

0.5

Hourly

W/m2

0

1,000

1

Every 15

humidity Air

Global radiation

minutes

Figure 34. Example of Comma-separated Value.

Time, 13:15, Temp, 25, Humidity, 45

120

Intelligent Agriculture

and re-route the data to coordinator, thereby acting as a bridge between the end devices and the coordinator for long-distance communication. A single Raspberry Pi 3 (acting as Super Master) when operating as a coordinator will receive and store CSV measurements from the routers in a CSV formatted file; every 24 hours, a new file will be created to store the new measurements. The saved CSV files are then opened via a computer and used to import the data into a structured query language (SQL) database to produce a graphical and tabular representation of the recorded data on a daily or weekly basis. Using text file in CSV format for the data sent from the end devices allows accessing the data in a simple manner, as well as to represent the data in tabular and graphical formats. This is due to its ability to have multiple recorded data consisting of one or more fields separated by commas; the comma is used as a field separator between different sources of data, such as time, temperature, and humidity. Therefore, having the comma separates the individually collected data into separate columns for a clear representation of the results. In addition, the use of a text file makes the results accessible over a wide range of database tools.

4.3.1. Measuring Outside Temperature and Humidity Measurement of the external factors that affect plant growth focus on features such as humidity and temperature. Humidity is the amount of water vapor in the air with respect to the maximum amount of water vapor the air can hold at certain temperatures. The stomata on the underside of plant leaves are affected by the changes in humidity, thereby impacting transpiration. At high humidity levels, plants have difficulty transpiring as there is a reduction in air circulation and plants cannot make water evaporate off their leaves or draw nutrients

Agrotechnology Solutions

121

from the soil. Prolonged high humidity levels can cause a plant to rot. At temperatures between 26°C and 32°C, a plant would close its stomata to reduce water loss, which is a cooling mechanism. Although the plant reduces its water loss, it also loses its ability to move carbon dioxide and oxygen molecules, which causes the plant to slowly suffocate. In the project, an HMP-60 sensor probe was connected to an Arduino Mega 2560 (acting as end device) to measure temperature and humidity and broadcast the measurements via an XBee PRO Series 1 to a Raspberry Pi (acting as router), saving the data into a text file with results in CSV format (see Table 19). The SW architecture to perform measurements using Arduino and transfer the collected measurements to Raspberry Pi is shown in Figure 35.

Table 19. Connections between HMP-60 Sensor and Arduino Mega 2560. Connections for HMP-60 Sensor Probe to Arduino Mega 2560

Wire

Wire Function

Color

Connection

Arduino

to Arduino

Function

Black

Temperature signal

A4

Analog input

White

Relative humidity signal

A5

Analog input

Blue

Power ground and signal reference

Vss

Ground

Brown

Power

Vcc

Power

Clear

Electric and magnetic field (EMF) shield

Vss

Ground

Intelligent Agriculture

122

Figure 35. Flowchart for HMP-60 Sensor Communication with Router. Start

Assign A5 to Analogpin

Assign A6 to Analogpin2

Assign Register 'Byte'as 0

NO

Temp & Humidity set as Input, Serial Began

Processed

Wait 2 Seconds

NO

YES Assign Temperature Value to Analogpin

Processed

YES Send Comma Separated Values of Temperature and Humidity

IS Temp Value and Humidity Value >0 YES

END

Is Byte Equal to 1 ?

YES NO

NO

Is Serial Available >0?

Convert Temperature and Humidity to Digital Value

Processed NO YES

YES

NO

Assign Humidity value to Analogpin2

Convert Digital Humidity to Humidity Value (a⫻100) – 5

Convert Digital Temp to Temperature Value (a⫻100) – 40

The flowchart for the Arduino SW architecture shows that the temperature and humidity connection of the HMP-60 sensor probe required an analog input. In this case, the Arduino receives the analog voltage and converts it into a digital value between 0 and 1,023. To get the measurements in degrees Celsius and humidity as a percentage, the digital values had to be converted back to analog and substracted from the error value, as described in the datasheet for the sensor. The conversion from digital to analog measured voltage

Agrotechnology Solutions

123

uses the system voltage as a maximum range and the resolution of the analog-to-digital conversion (ADC) of 1,023 (as shown by the example in Formula 1).

Analog voltage measured ¼

Sytem voltage × ADC value Resolution of ADC

ð1Þ

Example: System voltage = 5 V, ADC value = 434, and resolution of ADC = 1023. 5 × 434 ¼ 2:12V 1023 Once the conversion is completed, the humidity and temperature CSVs are ready to be sent to Arduino once it has received a value of “one” from the Raspberry Pi (RPi). The forthcoming paragraphs describe how the HMP-60 sensor is programmed to interface with the Arduino Mega 2560 to receive sensor measurements and transfer them to the Raspberry Pi. Figure 36 shows the pin assignments, configuration values, and configuration bits. The pin assignments are used to assign a name to each pin on the Arduino. Configuration values are used to assign a variable to zero so that new values can be stored in them: “TempVal” is used to store the input voltage value read from the temperature sensor, “HumidityVal” is used to store the input voltage value read from the humidity sensor, and “incomingByte” is used to temporally store the incoming value sent by the Raspberry Pi. The configuration bits are used to assign the HMP-60 connections to the Arduino as an input to receive the data from the sensor and setup a serial baud rate of

124

Intelligent Agriculture

Figure 36. Pin Assignments, Configuration Values, and Configuration Bits of HMP-60 Sensor.

9,600 for the rate at which data are transferred. Therefore, 9,600 bits per second are transferred. Figure 37 shows the conversion bits, configuration of analog inputs and reading from Raspberry Pi, and the sending of temperature and humidity values to Raspberry Pi. In the conversion bits, the term “float” is used to create floating point numbers that contain floating decimal points for computers to recognize real numbers. For both temperature and humidity, the voltage values read from the sensors are converted automatically by the Arduino as ADC, whereas for converting the voltage to temperature in degrees and humidity as a percentage, a digital-to-analog conversion (DAC) is required (example Formula 1). Once the voltage readings have been converted from digital to analog, the temperature can be obtained by multiplying the voltage value by a hundred and substracting 40 (see Figure 35).

Agrotechnology Solutions

125

Figure 37. Main Code of HMP-60 Sensor.

Analog inputs are configured to assign the voltage readings for temperature and humidity to the allocated analog input pins. Data are read from Raspberry Pi using “Serial. available()” to check infinitely using a “while loop” if the Arduino has received serial data on the universal asynchronous receiver/transmitter (UART) ports using “Serial.read ()”; if serial data are available, then it is stored in the incoming byte register. An “If” statement is used to check if the incoming byte is equal to the value of “one”; if the value is then greater than zero, then the data will be sent

126

Intelligent Agriculture

Figure 38. Raspberry Pi Software.

with CSVs using “Serial.print” so that the RPi can place them in the CSV file format. The last line of code uses the statement “Serial.println” so that when the next data line is sent, it will appear on a new line. Figure 38 shows the code on the Raspberry Pi to send requests to the Arduino and receive the temperature and humidity measurements. Serial input was setup by reading the USB port with the Xbee shield and assigning the same baud rate as the Arduino to receive the data on the same frequency range. The SW works very similar to the Arduino, the serial write sends a value of “one” to the Arduino to receive data every two minutes. The term “app” is assigned as an acronym to the CSV file and the data are saved into the CSV file by opening the file name “HMP60Trial.csv” and using acronym “a” to adjust the file. Then, the data are written into the file using “app.write,” and once the data have been written, it is closed using

Agrotechnology Solutions

127

“app.close.” The measurements are timestamped by using the RPi’s on-board clock. Figure 39 shows the point-to-point communication between the Arduino Mega 2560 (end device) and Raspberry Pi 3 (router) as a single node. From the figure it may be perceived that there is no data loss as the transferred temperature and humidity values have no gaps of error when received. Figure 40 graphically represents the measurements received using Microsoft Excel tool; the measurements were plotted against time that was set by the Raspberry Pi 3 (router) to receive data every two minutes. The data collected shows the temperature recorded as approximately 25.82°C with a humidity of approximately 48.4%.

4.3.2. Measuring Electrical Conductivity, Moisture, and Temperature of the Soil Measuring the EC (measured in milliSiemens per meter (mS/ m)) of the soil helps correlates the soil properties that affect crop productivity, such as the amount of moisture held by

Figure 39. Raspberry Pi, HMP-60 Received.

Intelligent Agriculture

128

Figure 40. Graphical Representation of HMP-60 Testing. Temperature & Humidity (ºC & %)

HMP-60 Sensor Test 60 50 40 30 20 10 0 15:07 15:08 15:10 15:11 15:12 15:14 15:15 15:17 15:18 Time Temperature

Humidity

the soil particles. Sandy soils have a low conductivity, silty soils have medium conductivity, and clayey soils have a high conductivity. This allows for irrigation management as the water content is measured using EC by estimating the amount of salts and iron available in the water. Moreover, the soil moisture level may be measured using EC with Formula 2. Measurement of soil temperature is important, as some temperatures can slow down plant growth and the germination of various seeds require different soil temperature ranges; therefore, temperature details will help monitor plant growth and identify field areas appropriate for the various types of seed plantation.  Soil moisture dS=m ¼ 4:3 × 106 × ε3a  5:5 × 104 × ε2a þ 2:92 × 102 × εa  5:3 × 102 ð2Þ

Agrotechnology Solutions

129

At this stage of testing, a 5TE sensor probe was connected to the Arduino and configured using SDI-12 serial interface. The sensor sends three values which are separated; the first value is the EC of the soil and the second value is a zero, which is ignored as it is used as a space between two values, and the final value is the temperature of the soil. Soil moisture level is obtained using Formula 2. The Raspberry Pi then receives these values wirelessly using Xbee. Distance was created between the Arduino and the Raspberry Pi by placing the former in another room; this helped to test the range of connection. Figure 41 shows the SW architecture for the 5TE sensor probe. The serial data received by the sensor are in alphanumeric format; therefore, conversion is done to enable the Arduino to read the values as digital data. Each measurement received is then allocated a name to understand what each value is for and another conversion is required to obtain soil moisture details (see Formula 2). Once all measurements are available, the Arduino sends the data in CSV format to be placed in a file on the RPi as database. The SW will then end and start again when requested by the RPi. The forthcoming paragraphs show how the 5TE sensor was programmed to interface with the Arduino Mega 2560 to receive sensor measurements and transfer them to the Raspberry Pi. Figure 42 contains the information for the header files, configuration values, and configuration bits for the 5TE soil sensor. Header files contain all the files to interface with sensor, and therefore containing the SDI-12 serial language which is then converted to readable data. Conversion values are used to set the maximum string length of data received from the sensor. Five samples are taken for each measurement because multiple readings help to generate average of a value to get better results; this is because noise or fluctuations

Intelligent Agriculture

130

Figure 41. Flowchart of Software for 5TE Sensor Probe. START

ASSIGN VARIABLES

ASSIGN DATA PIN TO PIN 2 OF ARDUINO

ASSIGN INPUT SAMPLES AND SDI-12 SERIAL TO DATA PIN

NO

PROCESSED

YES

SERIAL BEGIN

FLOAT VARIABLES

NO

PROCESSED

YES

NO CONVERT SDI-12 SERIAL DATA TO DIGITAL DATA

YES

GET MEASUREMENTS

PROCESSED

YES

IS SAMPLES DIELECTRIC MEAS 2 > IGNORE MEAS 3 > DEGREES

YES

CLEAR VALUE

YES CONVERT DIELECTRIC TO GET SOIL MOISTURE

YES

PROCESSED NO NO CLEAR MEMORY

IS CONVERSION DONE?

NO

YES

STORE ALL THREE MEASUREMENTS IN MEMORY REGISTER?

IS MEMORY CLEARED?

NO

PROCESSED

YES

SEND COMMA SEPARATED MEASUREMENT TO RPi END

may affect the measurements. Data cable of the sensor was connected to pin two of the Arduino and assigned as variable data for user-friendly reading. Two integers, “int,” were used to assign a one-second delay variable and the data location register. A serial baud rate of 9,600 was setup to give a data transfer rate of 9,600 bits per second.

Agrotechnology Solutions

131

Figure 42. Header Files, Configuration Values, and Configuration Bits of 5TE Sensor.

Figures 43 and 44 show the main code for the 5TE soil sensors which includes floating point number samples, samples obtained from sensor, conversions to retrieve other measurements, and measurements sent to RPi in CSV format. Floating point numbers were used in an array to hold the values of the five samples taken for the dielectric constant and temperature, and then in another array for obtaining the mean of each of the five measurement samples. This was then used to take an average value of each measurement to obtain a precise value of dielectric constant and temperature.

132

Intelligent Agriculture

Figure 43. Main Code for 5TE Soil Sensor (A).

Mathematically, this is done by finding the mean value of the five samples and then dividing the mean value by five to obtain its average. Samples are stored when string length, “strlen,” is greater than zero; “strtok” is used to separate the unwanted characters from the samples such as the SDI-12 characters ([]).

Agrotechnology Solutions

133

Figure 44. Main Code for 5TE Soil Sensor (B).

The statement “atof” was used to separate the decimal part in the the sample values from its integer part. A “while loop” was used to infinitely check if the Arduino has received serial data on its UART ports using “Serial.read()” and then storing

134

Intelligent Agriculture

that value in the “DataIn” register. An “if” statement is used to check if the serial data is equal to one; this helps to send the data to the Raspberry Pi through “Serial.print” statement. Function “get_measurement” is a character statement used to retrieve the SDI-12 serial data from the 5TE soil sensor. Figure 45 and 46 show the results gathered from a dry plant and one that has been watered; a difference between the two plant’s EC can be noticed as the values changed from 3.71 mS/ m to 16.31 mS/m. Therefore, the EC was high due to the amount of water held in the soil particles and also due to the soil tested had a content of clay. Also, a change was shown in the level of the moisture of the soil from 0.11 dS/m to 0.36 dS/m (equivalent 11 mS/m and 36 mS/m), which indicates that the plant had a little water content and changed to a high content when more salt and iron were added through watering. The measured temperature of the soil was high for both plants but reduced when water was added to the soil, as it helped in

Figure 45. Sensor Measurement of Dry Plant.

Agrotechnology Solutions

135

Figure 46. Sensor Measurement of Watered Plant.

cooling down the plant. High temperature was initially recorded because the plants stood in direct sunlight. Thereafter, the Arduino node was placed in another room with the sensor placed in the soil of a plant. Data were sent wirelessly through Xbee modules to a Raspberry Pi, where data were collated in CSV file format. The same code was used as before for the Raspberry Pi with only the file name changed to “5TETrial2.csv” (see Figure 47). Figures 47 and 48 show the data stored on the Raspberry Pi from the 5TE sensor on the Arduino. The results show that the soil temperature was approximately 24.6°C, dielectric constant was approximately 8.95 mS/m, and soil moisture level was 0.17 mS/m. Figure 45a shows that the data sent from the Arduino was sent with CSVs as expected.

Intelligent Agriculture

136

Figure 47. Raspberry Pi, 5TE Sensor Data Received.

Electrical Conductivity, Soil Moisture, and Soil Temperature

Figure 48. Graphical Representation of 5TE Sensor Data Received. 5TE Soil Sensor Test

30 25 20 15 10 5 0 14:09

14:11

14:12

14:13 14:15 Time

Electrical Conductivity

Soil Moisture

14:16

14:18

14:19

Soil Temperature

4.3.3. Measuring Global Radiation Global radiation is the total short-wave radiation from the sky falling onto a horizontal surface on the ground. It is

Agrotechnology Solutions

137

measured in watts per meter squared (Wm2). Monitoring global radiation facilitates managing the cropping processes such as optimum time for sowing, optimum plant population, timed application of fertilizers, and irrigation management. This helped planting certain plants for which the optimum global radiation increased their growth, resulting in more crop yield. At this stage of testing, an SP-212 global radiation sensor probe was connected to the Arduino through an analog input similar to the HMP-60 interface process requiring a DAC through Formula 1. The data were then sent to a RPi and stored in the database. Figure 49 shows the SW of the SP-212 sensor connected to an Arduino sending data to a RPi. The value read from the sensor is a voltage reading which gets converted to digital form using the Arduinos’ on-board ADC, which puts the value in a range between zero and 1,023. This value is then converted back by DAC. The read voltage value is then converted to W/m2 using the two mV per W/m2 conversion in the sensors’ datasheet. The forthcoming paragraphs show how the SPI-212 global radiation sensor was programmed to interface with the Arduino Mega 2560 to receive sensor measurements and transfer them to the Raspberry Pi. Figure 50 shows the configuration bits of the SPI-212 sensor. “Byte” was used as a register to hold the incoming values. A serial baud rate of 9,600 was setup to give a data transfer rate of 9,600 bits per second. Figure 51 shows the main code for the SPI-212 sensor. The sensor inputs an analog voltage value and the Arduino automatically converts it to a digital voltage in the range of zero to 1,023. Therefore, to convert the voltage to W/m2, the analog voltage was obtained by converting the digital voltage back to analog using Formula 1. To obtain voltage in W/m2, the SPI-212 datasheet was used. Floating point

Intelligent Agriculture

138

Figure 49. Flowchart of SPI-212 Global Radiation Sensor Arduino Software. START

ASSIGN INCOMING BYTE TO ANALOG PIN A0 OF ARDUINO

PROCESSED YES FLOAT VARIABLE

NO

PROCESSED YES DAC

IS VARIABLE CONVERTED? YES SEND COMMA SEPARATED VALUES TO RPi

END

NO

NO

Agrotechnology Solutions

139

Figure 50. Configuration Bits of SPI-212 Sensor Software on Arduino.

Figure 51. Main Code of SPI-212 Sensor Software on Arduino.

numbers, “float,” were used to use the formulas in a manner that the computer can recognize. “If” statements were used to check if the serial available is equal to “one” and less than 1,250 W/m2. If the serial statement is true, then the global radiation measurements would be transmitted to RPi in CSV format. An “else if” statement is used to transmit an error message if the incoming variable is greater

140

Intelligent Agriculture

Figure 52. Raspberry PI, SPI-212 Sensor Data Received.

than 1,250 W/m2. The range of global radiation was obtained from the SPI-212 sensor datasheet. Figure 52 shows the measurements gathered from the testing of the SP-212 sensor; the results show that at a higher intensity of light, the voltage was higher, whereas at a lower intensity of light, the voltage was lower. Figure 53 shows that the RPi received the data and stored it in a CSV database. The same code was used as before for the Raspberry Pi with only the file name changed to “SPI212.csv.” During the testing it was sunny and cloudy; when the sun shined more through the clouds, the

Agrotechnology Solutions

141

Figure 53. Graphic Representation of SPI-212 Sensor Data Received. Global Radiation Sensor Test

Global Radiation (W/m^2)

700 600 500 400 300 200 100 10 :5 10 3 : 10 55 : 10 57 :5 11 9 :0 11 1 : 11 03 : 11 05 :0 11 7 :0 11 9 : 11 11 : 11 13 : 11 15 : 11 17 : 11 19 : 11 21 :2 11 3 : 11 25 :2 11 7 : 11 29 :3 1

0 Time

voltage value represented in W/m2 increased (seen at 11.07 a.m. and 11.13 a.m.).

4.4. DEMONSTRATION OF SMART FARMING NETWORK In this chapter, the individual sensor parts described earlier were put together into a smaller scale of the network; therefore, the two routers were used to send data to a coordinator from the end devices connected to each individual router. This section describes how the end devices were setup with the routers and the SW that was involved (see Figure 54). The network was setup for demonstration purposes with two routers, with each router having end devices connected

Intelligent Agriculture

142

Figure 54. Small Scale of Network.

Data Transferred from Bridge-1 to Coordinator

Coordinator (Collecting Data from Both Bridges)

Data Transferred from Bridge-2 to Coordinator

Router (Bridge-1)

Router (Bridge-2)

Data Received from End Devices and Transferred to Bridge-2

Data Received from End Devices and Transferred to Bridge-1

Multiple End Devices with Sensors Connected to (bridge-2)

Multiple End Devices with Sensors Connected to (Bridge-1)

Figure 55. Example of Simple Bridge Connection. Router (RPi)

Coordinator (RPi) Data Received From End Device and Re-routed to Coordinator

End Device (Arduino) Data Transferred from End Device to Router

wirelessly using the Zigbee S2C modules. The purpose of each router was to act as a bridge between the coordinator and the end devices for expanding the range of data transfer as shown in Figure 55 as a single router and then two routers were used in the demonstration with multiple end devices.

Agrotechnology Solutions

143

In order to have two bridge networks, the destination addresses of the Zigbee S2C modules were changed because this address defines the destination the data would be transmitted to. This was achieved by setting the end devices’ Zigbee module “Destination Address High (DH)” value to that of the router’s “Serial Number High (SH)” value. For example, the SH value of the router is “13A200”; so the end device DH value needs to be set to “13A200” as well. The “Destination Address Low (DL)” value of the end device was also set to the router’s “Serial Number Low (SL).” For example, the routers SL value was “41502563,” so the DL value of the end device was also set to “41502563.” Therefore, this allowed the end devices connected to the first bridge to send data to that bridge only and not the second bridge. The same process was used for the second bridge but with different values of destination addresses. In order for the bridges to transfer data to the coordinator, the destination addresses DL and DH were set to the SH and SL values of the Coordinator Zigbee module; therefore DH = SH and DL = SL. The value of the addresses was obtained using Digi XCTU SW to read the Zigbee modules and write the addresses of DH and DL into the module.

4.4.1. Software Used during Network Testing This section describes the SW used for the coordinator, routers, and end devices during the testing of the small-scale smart farming network. The data were read from the Zigbee module connected to the Raspberry Pi’s USB port 0; a baud rate of 9,600 was used so that it had the same data transfer rate as the other Zigbee modules connected to the Routers, thereby allowing data to be received. Statement “DatIn” was used as a register to store the

144

Intelligent Agriculture

incoming data read from the USB port of the RPi using serial interface with “readline().” Statement “app” was used as a name to open, write, and close a CSV file named “SmartFarming5,” and the statement “a” was used to allow editing of the CSV file. Using the statement “app.write(DatIn + ”\n”)” allows the serial data received from the routers to be written into the CSV file to be stored as a database. Statement “app.close()” was used to close the CSV file once the incoming data had been written into the file to allow new incoming data to be written on the file. This process will continue to function while these statements are “True,” and therefore only when data are being received. Statements in light gray are irrelevant as they have been commented out (see Figure 56). The data were read from the Zigbee module connected to the Raspberry Pi’s USB port 0; a baud rate of 9,600 was used so that it had the same data transfer rate as the other Zigbee modules connected to the end devices, thereby allowing data to be received. Statement “IncomingData” was used as a register

Figure 56. Software on the Coordinator.

Agrotechnology Solutions

145

to store the incoming data read from the USB port of the RPi using serial interface with “readline().” Once the received data are stored in the register, the data are transferred to the coordinator using the statement “ser.write (IncomingData + ”\n”),” which writes the received data to the transmitting port (TX) of the Zigbee module. The statement “\n” is used to create a new line so that the data transferred is placed on a new line on the coordinator; this helps to visually represent the data in a clear manner (see Figure 57). The forthcoming paragraphs describe the SW used for transmitting the data from the sensors, as shown in Table 20. The SW used for obtaining the measurements from the sensor remained the same as described in the following text. Figure 58 shows the SW used for the 5TE sensor with the DHT22 sensor included. Statement “Serial.print” is used throughout to write all the data to the transmitting port (TX) of the Zigbee modules connected to the Arduino. A description of the node is provided so that the user is able to determine which data needs to fed to which node. The data are date and timestamped at the beginning to show when the data were collected; thereafter, soil dielectric constant (EC), soil moisture level, soil temperature, and internal temperature and humidity were measured. The internal temperature and humidity sensor is used to

Figure 57. Software on the Bridges.

Intelligent Agriculture

146

Table 20. Description of Devices Used. Sensors Used During Demonstration and Their Definitions Sensor

Definition

HMP-60

Measuring outdoor temperature and humidity

SPI-212

Measuring global radiation

5TE

Measuring soil temperature, soil moisture, and

DHT22

Measuring internal temperature and humidity of

electrical conductivity enclosures Timing Devices Used During Demonstration and There Definition Timing Device Definition DS3231

Real-time clock used for timestamping measurements

monitor the temperature and humidity within the enclosure; this processis undertaken as a safety protocol because the Arduino would need to be shut down if any moisture gets into the enclosure or the temperature reaches above 70°C. The statement “delay” is used to set the measuring frequency of the sensor; so for example, if the measurements are to be taken every 15 minutes, then the delay would be set to 900,000 milliseconds. For an upscale design of the network, the SW for the soil sensor (5TE) is just duplicated with a different measurement frequency. Figure 59 shows the SW used to send the data in CSV file format for the SPI-212 sensor with the DHT22 sensor included. Serial was used to write the data to the transmitting port (TX) of the Zigbee module connected to the Arduino. The date and time are also transmitted by using I2C protocol to retrieve the date and time from the real-time clock (RTC) through the serial data (SDA) and serial clock (SCL) ports. The internal

Agrotechnology Solutions

147

Figure 58. Data Send by 5TE Soil Sensor.

temperature and humidity details were retrieved from the DHT22 sensor using a serial interface for the enclosure properties mentioned earlier. A measurement frequency of seconds was used so that the global radiation is sent to the coordinator every two seconds for testing purposes in the laboratory. The data from the sensors will only be transmitted if the global radiation is not above 1,250 W/m2; if the value is higher, the sensor will fail and an “else if” statement is used to transmit an error message to the coordinator. Figure 60 shows the SW used to send the data in CSV file format for the HMP-60 sensor with the DHT22 sensor included. Serial statement “Serial.print()” was used to write the data to the transmitting port (TX) of the Zigbee module connected to the Arduino. The outdoor temperature and humidity details are only transferred to the bridge if the incoming temperature and humidity values are above zero.

148

Intelligent Agriculture

Figure 59. Data Send from SPI-212 Global Radiation Sensor.

Figure 61 shows the data stored on the coordinator, which is received from bridge one and two in CSV format, thereby allowing the data to be opened as a spreadsheet; a comma separates the data into separate columns for clear visual representation. The data can then be collected from the spreadsheet or a text file to build a SQL database. Figure 61 shows that each individual data has a description and has been date and timestamped for record purposes.

4.4.2. Graphical Representation of Results Collected from Each Sensor Used during the Demonstration In this section, 30 measurements were collected from the data collected for sensors HMP-60, SPI-212, 5TE, and DHT22.

Agrotechnology Solutions

149

Figure 60. Data Send from HMP-60 Outdoor Temperature and Humidity Sensor.

The measurements were then plotted on a graph against time showing variable changes when the sensors were tested in the laboratory. Therefore, the global radiation variable was changed by modifying the intensity of the lamp, outdoor temperature and humidity variables, which were changed by human contact. Additionally, the soil moisture, temperature and dielectric variables were also modified through human contact. Human contact refers to holding the sensors in one’s hands. Figure 62 shows the outdoor temperature and humidity sensor. The results show that the temperature recorded in the room was approximately 25°C and increased to approximately 30°C with human contact. Therefore, showing a change in temperature indicates that the sensor is operating as expected. However, the temperature only changes slightly, thereby showing that the sensor changes its variable

150

Figure 61. Data Collected on Coordinator.

Intelligent Agriculture

Agrotechnology Solutions

151

Figure 62. Graph of Outdoor Temperature vs Time.

gradually to the environment; this is useful for taking precise measurements. Figure 63 shows the outdoor temperature and humidity sensor. The results show that the humidity of the laboratory room was recorded at approximately 25.4%. With human contact, the humidity increased to approximately 88.9%. Therefore, this change in the humidity value of the sensor under contact indicates that the sensor is working as expected due to its adaptation to its current environment. The graph shows that the sensor adapts to the humidity within the environment rapidly. This is because the sensor has gaps with a filter within; by covering those gaps, air is trapped inside the sensor which is warmed up by the body temperature. Figure 64 shows the global radiation measurements for changing light intensity. The graph first shows a step change in light intensity which indicates that the Watts per meter measured increased rapidly with the change in light intensity. The maximum recorded global radiation was approximately 410 W/m2 and the minimum recorded was approximately 50 W/m2. The results indicate that the

152

Intelligent Agriculture

Figure 63. Graph Representing Outdoor Humidity vs Time.

sensor is sensitive to changes in light, which is useful for making accurate readings for outdoor global radiation as many a time there could be cloud coverage during the day, but the sensor will be capable of detecting light coming through between the clouds. Figure 65 illustrates the 5TE sensor showing results for soil dielectric constant measured against time. The results show that without human contact, the sensor detects the water vapor in the air to be approximately 1.12 mS/m. With human contact, the dielectric constant measured was approximately 11.7 mS/m, which indicates that the sensor has detected the water present on the hand. The change in the variable indicates that the sensor is sensitive enough to be capable of detecting even the water present on the hand. This sensitivity is useful for crop life as the sensor will be able to use the data regarding the water content in the soil to determine the amount of water in the area around the sensor (in meters). This will help determine the area around

Agrotechnology Solutions

153

Figure 64. Graph Representing Global Radiation.

the crops which is responsible for providing nutrients to the crop via water. Figure 66 displays the 5TE sensor showing results for soil moisture level measured against time. The results show that with human contact the level of moisture measured was approximately 0.3 deci-Siemens per meter (dS/m). The results indicate the sensor has detected the moisture present on the person’s hand; this shows the sensitivity of the sensor. Its sensitivity is useful for detecting the amount of moisture present in the soil, as at high temperatures the crop will begin to wilt; therefore, early detection of dry conditions will allow better crop growth. Figure 67 shows the 5TE sensor measurements for soil temperature against time. The results show that the sensor measured the room temperature of the laboratory at approximately 25°C, and with human contact, the measured temperature was 37°C, which is the normal core body temperature. This shows the sensor has high sensitivity which is useful for the crops, as it can detect high temperatures that may affect the water content in the soil.

154

Intelligent Agriculture

Figure 65. Graph Representing Soil Dielectric Constant vs Time.

Figure 66. Graph Representing Soil Moisture vs Time.

Agrotechnology Solutions

155

Figure 67. Graph Representing Soil Temperature vs Time.

Figure 68 shows the DHT22 sensor measurements for internal temperature and humidity. As the prototype was demonstrated without enclosures, the sensor measured the room temperature and humidity of the laboratory. The graph shows that the temperature was approximately 25°C and the humidity level was approximately 35%. 4.4.3. Software for Setting Real-time Clock This section explains how the SW for the RTC connected to each node was used to set the date and time for stamping the sensor measurements. The RTC used was a DS3231. Figure 69 shows the SW for the DS3231 RTC. The interintegrated circuit (I2C) pins’ data signal “SDA” and clock signal “SCL” are assigned as “rtc” to retrieve the data from the RTC. A serial baud rate of 9,600 was used to view the date and time on the serial monitor of the Arduino. The statement “rtc.begin ()” was used to open the inter-integrated circuit interface with

156

Intelligent Agriculture

Figure 68. Graph Representing Temperature and Humidity vs Time.

the RTC. In order for the Arduino and RTC to interface, the Arduino’s SDA was connected to the RTC’s SDA and the Arduino’s SCL was connected to the RTC’s SCL. The commented out statements “rtc.setDOW,” “rtc.setTime(16, 20, 0),” and “rtc.setDate(31, 8, 2018)” were used to set the day of the week, time, and date on the RTC’s on-board clock. Figure 70 shows the SW used for the DHT22 temperature and humidity sensor. The temperature and humidity data were received using single-wire protocol through pin 4 of the Arduino. The temperature is received from the negative temperature coefficient (NTC) thermistor of the sensor using the statement “dht.getTemperatureC().” The humidity is received using the statement “dht.getHumidity().” A delay of two seconds was used to view the measured temperature and humidity on the Arduino serial monitor every two seconds. In this section, the printed circuit board (PCB) for the connection between the sensors and Arduino is described. The sensors are connected to the Arduino using terminal blocks to allow

Agrotechnology Solutions

157

Figure 69. Software for Setting Time and Data on Real-time Clock.

easy adaption when the sensors need replacing. Female pin headers were used to connect the DHT22 and RTC as a plug-in function to allow easy replacement as well as to implement the Arduino using male to male jumper cables. Extra power and ground and sensor connections were added for any further adaption to any of the nodes. The printed circuit board dimensions were made to fit into the enclosures with ease to allow room for the Arduino and cables. Four mounting holes on each corner were added to mount the PCB into the enclosure so that connection is not lost during movement of the enclosure.

4.5. FEASIBILITY ANALYSIS INCLUDING PROTOCOL UPDATES AND PROOF TO CONTINUE WITH TRL4 AND BEYOND TRL is a metric-based system that assesses the program concepts, technological requirements, and technological capabilities.

158

Intelligent Agriculture

Figure 70. Software for DTH-22 Temperature and Humidity Sensor.

TRL is based on a scale from one to nine to show the maturity level of the research and technology development with nine being the most mature. The maturity level of the technology is assessed by device, material, component, and process. The TRL ratings help determine how far a particular technology is from being deployed by industry or public. Figure 71 demonstrates each requirement for the TRL. TRL0 and TRL1 are in the basic technological research development stage, with TRL0 being an unproven idea with no analysis or testing performed and TRL1 having observed and reported a published research document that underlines

Agrotechnology Solutions

159

Figure 71. Technology Readiness Levels.

the concept of the project. Research to prove feasibility is at TRL2 and TRL3 and part way to TRL4; at TRL2, the concepts have been formulated as a practical application with general assumptions and some analytical data to outline the concept reported at TRL1. TRL3 is the proof of concept with analytical and experimental studies performed on a laboratory scale to validate the

160

Intelligent Agriculture

predictions of the concept on a larger scale. Feasibility and technological demonstration falls within TRL4, as it is only a lowfidelity design due to being a laboratory-scale demonstration before the eventual system to establish whether the technological components will work together and obtain testing results. TRL5 is a laboratory-scale demonstration of the high-fidelity design with the technological components integrated and tested in a laboratory-simulated environment where the technology created is going to operate. System development falls between TRL6 and TRL7, with the prototype system designed and testing in an operational environment that is near or at a planned location of the system, integration and gathering data to compare with the simulated environment data. After testing at TRL6, the technological readiness changes to TRL7 if the operational environment demonstration was successful. The launch of the system and its operation fall between TRL8 and TRL 9. In this case, the fully completed and tested actual system has proven that the developed technology works in its final stage under the expected conditions. Then, after assessing the gathered data and ensuring it meets the operational requirements and then finalizing the design of the technology. At TRL9, the reporting of the performance of the system in its final design in an operational environment has been completed to start the process of putting the technology on the market. The proposed smart farming technology is currently at TRL3 as the concepts proved by taking the individual components of the eventual system and collating information. The process that followed to achieve the data for analytical prediction to the final concept was by programing the individual sensors and taking measurements of soil moisture, EC, soil temperature, outside temperature and humidity, and solar radiation. The collated data were used for analytical

Agrotechnology Solutions

161

predictions to validate that these sensors are operational and can be used to monitor the plant life and work with the chosen technology to retrieve the data (Arduino Mega 2560). The concept of wireless SNs was achieved by using two Xbee Pro’s, one connected to the Arduino and the other to the Raspberry Pi. This formed a wireless link, as RF was used to receive data from the Arduino and store it on the Raspberry Pi so that it can be used for a database. With TRL3 achieved, the proposed project can now move to TRL4 by connecting all the sensors to the Arduino and using a Raspberry Pi to collect the data from an Arduino. Therefore establishing the two technologies can work together and creating a single measurement node as part of the eventual design, as it is a low-fidelity design due to only having a single node whereas the eventual system has several nodes connected to a single Raspberry Pi. Adding more nodes will move the technology into TRL5 as it is the eventual design of the smart farming technology project.

This page intentionally left blank

CONCLUSIONS

This chapter presents a prototype system, demonstrating its functionality for retrieving data from sensors and relaying these data through a gateway or base station. Subsequently, the results are presented to users. The system features a custom sensor design for power efficiency, cost-effectiveness usage off the shelf, as well as scalability and ease of use. This design of the WSN system consists of a sensor node, Raspberry Pi as a base station, XBee as a networking protocol, and a number of open-source software packages. The major advantages of this system are low cost, low power consumption, compactness, scalability, easy deployment, and easy maintenance. Since the project is a prototype that was developed under some limitations and in a short time, there are certain tasks that should be conducted in the future, which would help develop the system further. • The communication applied must be replicated to the other pairs of Arduino and Raspberry Pi. • Knowledge of LabVIEW (Arduinomega, 2018), which is a visual programming tool for wiring together hardware devices installed on Raspberry Pi, must be acquired, which will enable sharing of variables. 163

164

Intelligent Agriculture

• Research is being conducted on how to create a MySQL (Mancuso and Bustaffa, 2006) database through Raspberry Pi and for putting Arduino in sleep mode, as it wakes up during data sampling and external interruptions. • A graphical user interface (GUI) must be developed to download parameters, initiate signals, acquire return transmissions, and display/format data. In this project, Raspberry Pi 3 Model B acts as a base station and two Arduino Mega 2560 boards act as end devices. In WSN, there are two types of devices: coordinators and end devices. There is only one coordinator in the network, which actually communicates with the base station. There can be more than one end devices. Experiments were conducted to establish a point-to-point communication between Arduino Mega 2560 and Raspberry Pi 3 Model B. Matplotlib software (Matplotlib, 2018) is used for graphical representation and for plotting the graphs. This software is used in Python scripts. It is a Python 2D plotting library which produces figures of publication quality in a variety of formats and interactive environments across platforms (Matplotlib, 2018). For experimental purpose, XBee is configured as a coordinator, which is connected with the Raspberry Pi. Other two XBee radios are configured as Router 1 and Router 2.

EXPERIMENTAL SETUP For the simulation, two potentiometers were connected to Arduino Mega 2560 (microcontroller A) to emulate analog sensors. One temperature sensor LM35 connected to Arduino Mega 2560 (microcontroller B) was used as an analog sensor. Then data were processed, and real-time data was sent to Raspberry Pi using the XBee radios (see Figure 72).

Conclusions

165

Figure 72. Smart Farming Wireless Sensor Network. (a) Raspberry Pi Connected to XBee. (b) Sensors Connected to Arduino Mega-A.

Figure 72a shows that the Raspberry Pi 3 Model B, which acts as a base station, is connected to XBee PRO Series 1 radio using USB. Figure 72b shows that the Arduino Mega 2560 (microcontroller A) is connected to XBee PRO Series 1 radio using XBee shield. Two potentiometers act as emulating analog sensors connected to A0 and A1 pins of Arduino Mega 2560 (A). Figure 73a shows that the Arduino Mega 2560 (microcontroller B) is connected to XBee PRO Series 1 radio using XBee shield. Temperature sensor LM355 is connected to analog pin A0. The LM35 temperature sensor is a precision integrated circuit (IC) temperature sensor. The output voltage of LM35 is directly proportional to the temperature. The LM35 does not need external calibration or trimming to provide accurate temperature range. It is a low-cost sensor. It has low output impedance and linear output. The operating temperature range for LM35 is 55° to +150°C (Gondchawar & Kawitkar, 2016). As is specified in ATmega2560 microcontroller datasheet of Atmel manufacturer, it incorporates a 10-bit analog-to-digital converter (ADC). With the rise in temperature, the output

Intelligent Agriculture

166

Figure 73. Smart Farming Wireless Sensor Network Communication. (a) Temperature Sensor LM35 Connected to Arduino Mega-B. (b) Construction of a Wireless Sensor Network.

voltage of the sensor increases linearly; the voltage value is fed to the microcontroller, which multiplies it with the conversion factor to generate the actual temperature. Figure 73b shows the construction of a sensor node prototype and a WSN prototype using development boards for sensor and data node sets of sensors for measuring relevant field variables, wireless communication daughter boards. One XBee is configured as coordinator and is connected with the Raspberry Pi. The other two XBee radios are configured as Router 1 and Router 2. Router 1 sends LM35 (Aqeel-urRehman et al., 2018) temperature values and Router 2 sends two potentiometer values to the Raspberry Pi.

RESULTS Figure 74 shows the serial monitor of Arduino Mega 2560. Analog sensors, such as temperature sensor LM35, are used to read values from the sensor node by using Raspberry Pi. At the maximum temperature, the output will be 1 V and the maximum value of ADC will be 5 V;

Conclusions

167

Figure 74. Arduino Mega 2560 Serial Monitor-A.

this means that a percentage of the range is lost. It is possible to switch the internal voltage reference of a microcontroller to 1.1 V by a code. Sensors are used to expand the capabilities of the Arduino. Sensor output is connected to the input pin of Arduino and the data are converted to digital form. Some sensors have an analog-to-digital converter embedded to the sensor so that the data output is as digital data. In those sensors that don’t have on-board analog-todigital converter, data are sent in analog mode to Arduino, which then uses its on-board converter to convert the data to digital format. After the data are processed to digital form, it can be processed further on the microcontroller. A one-to-one connection between Arduino Mega 2560 A (runs C code) to Raspberry Pi 3 (runs python) was successfully achieved using XBee radios which are configured in application programming interface (API) and attention (AT) modes. Figures 74 and 75 show the temperature reading from the sensor node, which is plotted graphically using the software Matplotlib (Matplotlib, 2018). Figure 76a shows the Arduino Mega 2560 (B) serial monitor readings. Analog sensors are used to read values from the

Intelligent Agriculture

168

Figure 75. Graphical Representation in Raspberry Pi (Python Shell). Temperature from Arduino

45

45

40

40

35

35

30 25

30 25 20

20

15

15 10

Temperature from Arduino

50

Temp ºC

Temp ºC

50

0

5

10 15 Timescale

20

25

10

0

5

10 15 Timescale

20

25

Figure 76. Sensor Reading from the Sensor Node Plotted Graphically. (a) Arduino Mega Serial Monitor. (b) Raspberry Pi  Python Shell.

sensor node. The microcontroller has a 10-bit ADC; therefore, the values of the pin where the potentiometer sensors are connected will be between 0 and 1,024. So, with an output of 5 V, the ADC value will be 1,024, and with an output of 0 V, the ADC value will be 0. Figure 76b shows that Arduino B sends data to Raspberry Pi wirelessly, which is

Conclusions

169

represented graphically by using Matplotlib software (Matplotlib, 2018). Point-to-point wireless communication was successfully established between two Arduino Mega boards and one Raspberry Pi to create a prototype of a wireless sensor network.

This page intentionally left blank

REFERENCES Afram, A., & Janabi-Sharifi, F. 2014. Theory and applications of HVAC control systems: a review of model predictive control (MPC), Building Environment, 72(1), 343355. Aguilar, J.V. 2016. Predictive control of irrigations canals: robust design and real-time implementation, Water Resource Management, 30(11), 38293843. Akyildiz, I.F., Su, W., Sankarasubramaniam, Y., & Cayirci, E. 2002. A survey on sensor networks, IEEE Communications Magazine, 40(1), 102114. Arduinomega. 2018. Webpage. Available at: http://www.egr. msu.edu/classes/ece480/capstone/spring14/group08/ MaramAN.pdf [Accessed 12 March 2018]. Blackmore, S. 1994. Precision farming: an introduction, Outlook on Agriculture Journal, 23(1), 275280. Bogue, R. 2017. Sensors key to advances in precision agriculture. Sensor Review, 37(1), 16. Bumroongsri, P., & Kheawhom, S. 2014. Off-line robust constrained mpc for linear time-varying systems with persistent disturbances, Mathematics Problems & Engineering, 1(1), 282290.

171

172

References

Chien-Hsing, W., Shu-Chen, K., Yann-Yean, S., & ChuanChun, W. 2005. Targeting customers via discovery knowledge for the insurance industry, Export Systems with Applications, 29(1), 291299. Christofides, P.D. 2013. Distributed model predictive control: a tutorial review and future research directions, Computing Chemical & Engineering, 51(14), 2141. Corke, P., Wark, T., Jurdak, R., Hu, W., Valencia, P., & Moore, D. 2018. Environmental wireless sensor networks, Proceedings of the IEEE, 98(11), 19031917. Delgoda, D. 2016. Irrigation control based on model predictive control (MPC): formulation of theory and validation using weather forecast data and AQUACROP model. Environment Modelling Software, 78(1), 4053. Ding, Y., Wang, L., Li, Y., & Li, D. 2018. Model predictive control and its application in agriculture: a review, Computers and Electronics in Agriculture, 151(1), 104117. Donato, J.M., Schryver, J.C., Hinkel, G.C., Schmoyer, R.L., Leuze, M.R., & Grandy, N.W. 1999. Mining multidimensional data for decision support, Future Generation Computer Systems, 15(3), 433441. Farhadi, A., & Khodabandehlou, A. 2016. Distributed model predictive control with hierarchical architecture for communication: application in automated irrigation channels, International Journal of Control, 70(8), 5969. Fayyad, U., & Stolorz, P. 1997. Data mining and KDD: promise and challenge, Future Generation Computer Systems, 13(2/3), 99115.

References

173

Feelders, A., Daniels, M., & Holsheimer, M. 2000. Methodological and practical aspects of data mining, Information & Management, 37(1), 271281. Fengyuan, R., Haining, H., & Chuang, L. 2003. Wireless sensor network, Journal of Software, 14(1), 12821291. Figueiredo, J. 2013. SCADA system with predictive controller applied to irrigation canals, Control Engineering Practice, 21(6), 870886. Garcia, C.E., & Morari, M. 1982. Internal model control: a unifying review and some new results, Industrial Engineering & Chemical Process Designed Development, 21(2), 308323. Garriga, J.L., & Soroush, M. 2010. Model predictive control tuning methods: a review, Industrial Engineering & Chemical Responsibility, 49(8), 35053515. Gondchawar, N., & Kawitkar, R. S. (2016). IoT based smart agriculture. International Journal of Advanced Research in Computer and Communication Engineering, 5(6), 838–842. Han, J., & Fu, Y. 1999. Mining multiple-level association rules in large databases, IEEE Transactions on Knowledge and Data Engineering, 11(5), 798805. Horváth, K. 2015. New offset-free method for model predictive control of open channels, Control Engineering Practices, 41(1), 1325. Jaber, A. 2017. Design of an Intelligent Embedded System for Condition Monitoring of an Industrial Robot, Cham, Springer International Publishing. Jin-Shyan, L., Yu-Wei, S., & Chung-Chou, S. 2018. A comparative study of wireless protocols: bluetooth, UWB, ZigBee, and Wi-Fi. Paper presented at the 33rd Annual

174

References

Conference of the IEEE Industrial Electronics Society (IECON), Taipei, Taiwan. Kamilaris, A. 2018. Deep learning in agriculture: a survey, Computing & Electronics Agricultural, 147(1), 7090. Kohavi, R., & Provost, F. 2001. Data mining and knowledge engineering, IEEE Transactions on Knowledge and Data Engineering, 5(1/2), 1122. Lahiri, K., Raghunathan, A., Dey, S., & Panigrahi, D. 2002. Battery-driven system design: a new frontier in low power design. Proceedings of ASP-DAC/VLSI Design 2002. 7th Asia and South Pacific Design Automation Conference and 15th International Conference on VLSI Design, Bangalore, India, 11–11 January. Lal, R. 1990. Soil erosion and land degradation: the global risks, New York, NY, Springer. Lloyd, C. 2017. High resolution global gridded data for use in population studies, Science Data, 4(1), 1725. Mancuso, M., & Bustaffa, F. (2006). A wireless sensors network for monitoring environmental variables in a tomato greenhouse. Proceedings of the 2006 IEEE International Workshop on Factory Communication Systems, Torino, Italy, 28–30 June. Matplotlib. 2018. Installation — Matplotlib 2.2.0 documentation. Available at: https://matplotlib.org/ [Accessed 12 March 2018]. McCarthy, A.C. 2014. Simulation of irrigation control strategies for cotton using model predictive control within the VARIwise simulation framework, Computing & Electronics Agricultural, 101(1), 135147.

References

175

Methley, S., Forster, C., Gratton, C., Bhatti, S., & Teh, N. J. (2008). Wireless Sensor Networks Final Report, Version 2, Plextek. Available at: https://www.ofcom.org.uk/__data/ assets/pdf_file/0027/27387/wsn1.pdf [Accessed 11 April 2019]. Ncalculators. 2018. mAh battery life calculator & calculation. Available at https://ncalculators.com/electrical/ battery-life-calculator.htm [Accessed 12 March 2018]. Negenborn, R.R. 2009. A non-iterative cascade predictive control approach for control of irrigation canals. Proceedings of the IEEE International Conference on Systems, Man and Cybernetics, pp. 35523557, San Antonio, TX. Pitta, D. 1998. Marketing one-to-one and its dependence on knowledge discovery in databases, Journal of Consumer Marketing, 15(5), 468480. Pomeranz, K. 2009. The great Himalayan watershed: water shortages, mega-projects and environmental politics in China, India and southwest Asia, Asia-Pacific Journal, 5(1), 1122. Powers, S. (2012). The open-source classroom: your first bite of Raspberry Pi. Linux Journal, 224(7). Qin, S.J., & Badgwell, T.A. 2003. A survey of industrial model predictive control technology, Control Engineering Practices, 11(7), 733764. Rakhmatov, D., Vrudhula, S., & Wallach, D. 2003. A model for battery lifetime analysis for organizing applications on a pocket computer, IEEE TVLSI, 11(1), 10191030. RaspberryPi. 2018. Webpage. Available at: http://en. wikipedia.org/ Rehman, A. U., Abbasi, A. Z., Islam, N., & Shaikh, Z. A. (2014). A review of wireless sensors and networks’

176

References

applications in agriculture. Computer Standards & Interfaces, 36, 263–270. Roca, L. 2016. Predictive control applied to a solar desalination plant connected to a greenhouse with daily variation of irrigation water demand, Energies Journal, 9(3), 194211. Salahou, M.K. 2013. Research journal of applied science engineering and technology: review article control of an irrigation canals, Research Journal of Applied Science and Technology, 5(15), 39163924. Sanchez, P.A. 2002. Ecology: soil fertility and hunger in Africa, Science Journal, 29(5), 5562. Scattolini, R. 2009. Architectures for distributed and hierarchical model predictive control: a review, Journal of Process Control, 19(5), 723731. Shapiro, R.J. 2014. The U.S. software industry: an engine for growth and employment, New York, NY, SIIA. Sheikh, F., & Xinrong, L. 2018. Wireless sensor network system design using Raspberry Pi and Arduino for environmental monitoring applications. Presented at the Elsevier 9th International Conference on Future Networks and Communications. Shim, J.P., Wakenting, M., Courtney, J.F., Power, D., Sharda, R., & Carlsson, C. (2002). Past, present, and future of decision support technology, Decision Support Systems, 33(1), 111126. Sung, H.H., & Sang, C.P. 1998. Application of data mining tools to hotel data mart on the intranet for database marketing, Expert Systems with Applications, 15(1), 131. Tian, Y., Lv, Y., & Tong, L. 2013. Design and application of sink node for wireless sensor network, COMPEL: The

References

177

International Journal for Computation and Mathematics in Electrical and Electronic Engineering, 32(2), 531544. Voltaicsystems. 2018a. Arduino. Available at: https://www. voltaicsystems.com/solar-arduino-guide [Accessed 12 March 2018]. Voltaicsystems. 2018b. 6 Watt solar kit. Available at: https:// www.voltaicsystems.com/6-watt-kit [Accessed 12 March 2018]. Vukov, M. 2015. Real-time nonlinear MPC and MHE for a large-scale mechatronic application. Control Engineering Practices, 45(1), 6478. Wang, Y.G. 2001. PID auto tuner and its application in HVAC systems, American Control Conference, 3(2), 21922196. Werner-Allen, G., Lorincz, K., Ruiz, M., Marcillo, O., Welsh, M., Johnson, J., & Lees, J. 2006. Deploying a wireless sensor network on an active volcano, IEEE Internet Computing, 10(1), 1825. Wu, B.S. 2008. Automatic canal control system and its operation and design, Advanced Water Science, 19(5), 746755. Wu, C.H. 2003. Data mining applied to material acquisition budget allocation for libraries: Design and development, Expert Systems with Applications, 25(3), 401411. Zhang, J. 2017. Robust model predictive control of the automatic operation boats for aquaculture, Computing & Electronics Agricultural, 142(1), 118125. ZigBee. 2006. Specification ZigBee Alliance. Available at: http://www.zigbee.org/

This page intentionally left blank

INDEX AgDrone System™, 910 Agricultural machinery, 20 Agrochemicals, 21, 2527 advantages of, 2526 disadvantages of, 2627 Agroptima, 12, 14 Agrotechnology, 8183 Agrotechnology solutions, 111116 aim and objectives, 116117 feasibility analysis, 157161 proof-of-concept activities, 118141 Agrowin, 11, 14 Aguascalientes, 35 Aguasmarket, 12, 14 Arduino Mega 2560, 59, 67, 7475, 7677, 118119, 122124, 126127, 129130, 133134, 136140, 145147, 155157, 160161, 163, 164, 166169 hardware components, 73 technical specification, 78

Artificial neural networks, xviii ATmega2560 microcontroller, 165166 Atmospheric sensors, technical specification of, 62, 63 Battery life, 6970 Battery power consumption calculation, 71 BIS Research, 5859 Brioagrow, 89, 14 Broccoli (Brassica oleracea) plantation of, 47 producers, in Mexico, 4447 Business user case, 8486 Campeche, 35 Cargill, 2021 Chiapas, 36 Coahuila, 36 Colima, 36

179

180

Comma-separated value (CSV), 119120, 123, 125127, 129, 131132, 143144 Crop menu sequence, 99 Crop user case description, 8788 Data acquisition (DAQ) system, 114115, 117 Database (DB) creation, 13 development and implementation of, 34 sequence, filling, 99 user case description, filling, 90, 92, 96 Data mining, 13 DHT22 sensor, 146147, 155, 156157 Differentiated integral processes, xviii Direct sowing, 21, 2425 advantages of, 24 disadvantages of, 2425 DS3231 RTC, 155156 Durango, 37 EC-05 (soil moisture sensor), 6165 Electrical conductivity (EC) measurement, 127135 Employee administration user case description, 92, 97

Index

Employee management sequence, 99 Entity-relationship diagram (ERD), 98, 101 5TE sensor probe, 129130, 131132, 133134, 135, 145146, 152153 5TM (soil temperature and moisture sensor), 6061 Fuzzy logic, xviii Geographical information system (GIS), 2223 Global positioning system (GPS), 2223, 5354, 5859 Global radiation measurement, 136141 Graphical user interface (GUI), 164 GS3 (electrical conductivity sensor), 61 Guanajuato, 37 Guerrero, 3738 Hardware brain, 7380 Arduino Mega 2560, 7677, 78 Raspberry Pi 3 Model B, 7476 XBee PRO Series 1, 7780 HCI, 99101 Hidalgo, 38 Hierarchical control model, 5051

Index

Historic data use case diagram, 8687 HMP-60 (humidity and temperature probe sensor) sensor, 60, 121, 136137 configuration values, 123124 HoneyComb, 910, 14 Humidity measurement, 120127 Hybrid seeds, 20 IEEE 802.15.4 standard XBee PRO Series 1, 78, 7980 Intellectual property analysis, 2833 computer programs, protection of, 3033 Internal control models, xxxxi Jalisco, 38 LM355 sensor, 165 MakeMap™ Processing, 910 Matplotlib software, 164, 167169 Monsanto, 2021 Morelos, 38 Nayarit, 39 Non-linear control systems, 4950 Nuevo León, 39

181

Oaxaca, 39 Optimal control systems, 4950 Outside temperature measurement, 120127 Planting techniques, 2122 Prediction models, xixxx, xx, xxixxii Proagro Productivo, 34 Production control, xixxx, xx, xxixxii Proof-of-concept activities, 118141 Puebla, 39 Qualcomm, 3031 Querétaro, 39 Quintana Roo, 3940 Raspberry Pi 3 Model B, 67, 7476, 118120, 123124, 125127, 129, 133134, 135, 137, 140141, 143145, 160161, 163, 164, 166169 hardware components, 73 software, 126127 technical specification of, 75 Remote control agriculture, 22 Sagarpa, 3442 San Luis Potosí, 40

182

Sensor nodes (SNs), 114, 115116, 117 Sensors, 2223, 5565 atmospheric sensors, technical specification of, 62, 63 EC-05 (soil moisture sensor), 6165 5TM (soil temperature and moisture sensor), 6061 GS3 (electrical conductivity sensor), 61 HMP-60 (humidity and temperature probe sensor), 60 sequence management, 99 soil sensors, technical specification of, 6265 SP-212 (solar radiation sensor), 60 use case description, management of, 8990, 95 Sinaloa, 40 Sink nodes, 6567 structure of, 6566 Sismagro, 1112, 14 Smart Agriculture, xviii Smart Fertilizer Management Software, 1011, 14 Smart monitoring and control system (SMCS), 99101, 111112, 114116

Index

Soil compaction, 16, 17 erosion, 16 moisture, 127129 productivity improvement techniques, 16 sensors, technical specification of, 6265 temperature measurement, 127135 Solar panels, 6773 battery power consumption calculation, 71 technical specification of, 72 V44 batteries, 72 Sonora, 40 SPI-212 sensor, 60, 137140, 146147 System architecture, 4952 hardware brain, 7380 Arduino Mega 2560, 7677, 78 components, 73 Raspberry Pi 3 Model B, 7476 XBee PRO Series 1, 7780 wireless sensor network, 5273 sensors, 5565 sink nodes, 6567 solar panels, 6773

Index

Tabasco, 41 Tamaulipas, 41 Technology readiness level (TRL), 115116, 157161 Tlaxcala, 41 Transgenic seeds, 2021, 2728 advantages of, 2728 disadvantages of, 28 UAA, 115116 University of Derby (UoD), 115116 V44 batteries Arduino runtime from, 72 technical specification of, 72 VegHands, 78, 14 Veracruz, 4142 Volumetric water content (VWC), 6061

183

description and functions, 85 sensors, 5565 EC-05 (soil moisture sensor), 6165 5TM (soil temperature and moisture sensor), 6061 GS3 (electrical conductivity sensor), 61 HMP-60 (humidity and temperature probe sensor), 60 SP-212 (solar radiation sensor), 60 structure of, 54 system architecture of, 55 XBee PRO Series 1, 59, 67, 7780, 121, 163, 164, 165, 166, 167 hardware components, 73 Yucatán, 42

Water conservation, 111 Wireless sensor network (WSN), 5273, 84, 113114, 115118, 163, 164, 166

Zacatecas, 42 Zigbee network, 118 Zigbee S2C modules, 141142, 143