Industry 4.0 Challenges in Smart Cities (EAI/Springer Innovations in Communication and Computing) 3030929671, 9783030929671

Citation preview

EAI/Springer Innovations in Communication and Computing

Dagmar Cagáňová Natália Horňáková   Editors

Industry 4.0 Challenges in Smart Cities

EAI/Springer Innovations in Communication and Computing Series Editor Imrich Chlamtac, European Alliance for Innovation, Ghent, Belgium

The impact of information technologies is creating a new world not yet fully understood. The extent and speed of economic, lifestyle and social changes already perceived in everyday life is hard to estimate without understanding the technological driving forces behind it. This series presents contributed volumes featuring the latest research and development in various information engineering technologies that play a key role in this process. The range of topics, focusing primarily on communications and computing engineering include, but are not limited to, wireless networks; mobile communication; design and learning; gaming; interaction; ehealth and pervasive healthcare; energy management; smart grids; internet of things; cognitive radio networks; computation; cloud computing; ubiquitous connectivity, and in mode general smart living, smart cities, Internet of Things and more. The series publishes a combination of expanded papers selected from hosted and sponsored European Alliance for Innovation (EAI) conferences that present cutting edge global research as well as provide new perspectives on traditional related engineering fields. This content, complemented with open calls for contribution of book titles and individual chapters, together maintain Springer’s and EAI’s high standards of academic excellence. The audience for the books consists of researchers, industry professionals, advanced level students as well as practitioners in related fields of activity include information and communication specialists, security experts, economists, urban planners, doctors, and in general representatives in all those walks of life affected ad contributing to the information revolution. Indexing: This series is indexed in Scopus, Ei Compendex, and zbMATH. About EAI - EAI is a grassroots member organization initiated through cooperation between businesses, public, private and government organizations to address the global challenges of Europe’s future competitiveness and link the European Research community with its counterparts around the globe. EAI reaches out to hundreds of thousands of individual subscribers on all continents and collaborates with an institutional member base including Fortune 500 companies, government organizations, and educational institutions, provide a free research and innovation platform. Through its open free membership model EAI promotes a new research and innovation culture based on collaboration, connectivity and recognition of excellence by community.

Dagmar Cagáˇnová • Natália Horˇnáková Editors

Industry 4.0 Challenges in Smart Cities

Editors Dagmar Cagáˇnová Slovak University of Technology in Bratislava Slovakia

Natália Horˇnáková Slovak University of Technology in Bratislava Slovakia

ISSN 2522-8595 ISSN 2522-8609 (electronic) EAI/Springer Innovations in Communication and Computing ISBN 978-3-030-92967-1 ISBN 978-3-030-92968-8 (eBook) https://doi.org/10.1007/978-3-030-92968-8 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023, corrected publication 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

This book deals with the next level of innovative technologies influencing industry and connectivity sectors in the future industrial, urban, and sustainable development. It provides a platform for synergy of ideas within smart industrial innovations, approaching them from various points of view: industry and management 4.0, expansion of new business models, smart technologies identifying up-to-date global challenges, new trends and opportunities. New managerial ideas, cutting-edge innovations, and technologies for mobility agenda are highlighted together with a multidisciplinary perspective approach. This unique point of view will inspire researchers, graduate students, and those who are interested in Smart City topics. Readers will find the answers for various questions connected to global and interconnected modern society. The book comprises the following 12 chapters: Chapter 1 investigates the ways to safe implementation and authentication for IoT devices, tying together all associated algorithms, protocols, and network systems. A number of techniques to address related challenges are proposed, such as elliptic curve cryptography (ECC), quantum cryptography, public keys, private keys, pseudonymous certificates, and more. Moreover, a simulation has been carried out to show the efficiency and security of a specific protocol, which proved to be powerful against de-synchronization, replay, and session hijacking attacks. Chapter 2 is oriented towards determining the best logistical route for the trade of goods between China and Ukraine as a mathematical model. Four different options and their related infrastructure were considered—railway, maritime, road, and air transport—in an experiment. For each option, order delivery volumes of the corresponding type of goods, cargo transportation volume in the current batch, and risk assessment factors using corresponding kinds of transport were evaluated to obtain the final profits of enterprises. Chapter 3 regards the SIM evolution, the main variances of M2M eSIM networks and approaches to their commercialization. Basic information about architecture, interfaces, and operations is mentioned, as well as three types of M2M eSIM networks—Operator-controlled variance, OEM-controlled variance, and M2M service provider-controlled variance. Before choosing the commercialization approach, v

vi

Preface

leading ecosystem players are to consider available resources, limitations, advantages, and disadvantages of each option. M2M eSIM technology adoption is still posed with the challenge of complex mesh networks and a long list of standardized and non-standardized interfaces, also acting as an opportunity for future research. Chapter 4 puts focus on biodegradable municipal waste, and links it to the key issue of most cities in the world concerning energy sources: those being insufficient energy independence, increased transportation costs, and high carbon footprint. By burning the raw materials present in municipal wastes—of which the majority is their biodegradable component, making it a clean, renewable resource—a burden is lifted off city landfills and a source of clean, environmentally friendly alternative fuel is obtained. Benefits of this closed cycle approach are examined in relation to city parks, gardens, or in the surrounding fields and forests near cities. Chapter 5 delves into the problem of public transport overcrowding, affecting not only the passengers’ comfort, but also raising efficiency and financial issues. An embedded system is proposed to count people in the means of public transport, combining a variety of infrared sensors for object detection, detection of direction, and object positioning. Multiple controlled experiments were carried out with a satisfactory success rate of up to 95%. Chapter 6 examines the cultural evolution towards the Fourth Industrial Revolution, bringing about changes in the fields of economic, social, political, and cultural capital. The basic attribute of the human race is culture, and the human ability to transfer artefacts, cultural technology, and findings cumulatively in time allows for its constant growth and continuity. Today’s digital economy is faced with a question of how the quality of human life might change under the determining influence of Industry 4.0. Chapter 7 proposes a case study from Slovakia regarding sustainable urban mobility in hopes for greener cities. The levels of dust, noise, and air pollution from the traffic dropped significantly as a result of the COVID-19 pandemic. Even though app-based and shared-ride services have risen in popularity, individual car transport dominates at the expense of sustainable modes of transport in most of Slovakia, leading to traffic congestion during peak hours. The possibilities of multimodality are shown in the context of short distance moves in the city of Nitra through a conducted marketing research. The findings show significant contrast in residents’ attitudes in different urban areas. Chapter 8 also presents a case study concerning the effects of the COVID-19 pandemic, specifically in city transport systems in the Slovak Republic. The results prove that the measures taken by the government have strongly minimized the impacts of the virus, flattening the curve swiftly. This chapter also zooms in on additional solutions to help fight the spread of the virus and improve passengers’ safety in transport systems, describing the concepts of Smart City, Safe City, smart transport and related systems. The focus of the proposed measures is given to the original design of mass transport system. Chapter 9 considers road detection as a means for minimizing health hazards while driving. It states that currently, road condition is monitored infrequently due to it being both time consuming and costly. Specifics of a developed IoT system

Preface

vii

is examined in the chapter in order to crowdsource a system ensuring the health of roadways by informing transit authorities of pothole locations, speeding up the whole monitoring process. Such a system includes smart sensors and performs machine learning on accelerometer data, processes and analyses acquired data without using the cloud, and sends it to a web server, sharing the location of road hazard with the responsible transit authorities. Chapter 10 deals with the newly arisen digitalization requirements and legal obligations for carrier companies caused by COVID-19. A concrete example of a food retailer’s changed workflow is examined as a result of having to comply with these regulations, using digital instruments as one of the key tools to keep their distance. The advantages and disadvantages of digital technology in the transport industry are considered, and it is shown how these technologies reduce certain types of costs and increase efficiency and transparency. Chapter 11 reviews the possibility of Czech countries to enter the Smart Factory revolution. Three factors are considered—point of departure, access to new technologies, and realization and successfulness—while assessing the reason for Czech countries success in adapting to all the industrial revolutions throughout the past, as well as mentioning the importance of geopolitics and international policy. For the main research method, description, analysis, and comparison are used. Chapter 12 presents the results of analysis of 50 IoT privacy policies in order to determine whether IoT manufacturers collect personal information about their users. The need for this method arose with the increased popularity of IoT devices, where the privacy policy statements are too long and too complicated, leaving the user confused and often letting the device access sensitive information without even realizing it. The used method studies the complicated and ambiguous statements in-depth, mimicking how an ordinary person reads and understands such policies sentence by sentence, even categorizing and labelling personal information according to its sensitivity level with a supervised machine. The high accuracy achieved by the classifier (98.8%) proves its reliability and validity. Lastly, the editors would like to express their sincere thanks to the authors of the chapters for contributing their outstanding knowledge, experience, and latest research results towards the creation of this book. Trnava, Slovakia

Dagmar Cagáˇnová Natália Horˇnáková

Contents

Secure Algorithm for IoT Devices Authentication. . . . . . . . . . . . . . . . . . . . . . . . . . . . Vincent Omollo Nyangaresi, Anthony J. Rodrigues, and Silvance O. Abeka Mathematical Modeling as a Tool for Selecting a Rational Logistical Route in Multimodal Transport Systems . . . . . . . . . . . . . . . . . . . . . . . . . . Olexiy Pavlenko, Dmitriy Muzylyov, Natalya Shramenko, Dagmar Cagáˇnová, and Vitalii Ivanov

1

23

Commercializing M2M eSIM Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bassem Ali Abdou

39

Closed Cycle of Biodegradable Wastes in Smart Cities . . . . . . . . . . . . . . . . . . . . . . Michal Holubˇcík, Jozef Jandaˇcka, and Juraj Trnka

59

Automated People Counting in Public Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peter Pištek, Simon Harvan, and Michal Valicek

75

Industry 4.0: From Smart Factories to Artificial Intelligence. . . . . . . . . . . . . . . Václav Soukup

95

Sustainable Urban Mobility–Multimodality as a Chance for Greener Cities: Evidence from Slovakia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Mária Holotová, L’udmila Nagyová, Tomáš Holota, and Dagmar Cagáˇnová Transport System of the Smart City Concept vs Pandemics: COVID-19 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Maroš Lacinák and Jozef Ristvej Drive Health: Road Condition Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Peter Ferguson, Brian Walker, Navid Shaghaghi, and Behnam Dezfouli Digitalization in Transport Logistics due to COVID-19: A Case Study from Germany . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Zuzana Papulová, Christian Korge, and Stephan Pritzl

ix

x

Contents

Smart Factory as the Top of the Development of the Industrial Revolution in Czech Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Vít Rouˇc and Štˇepán Strnad Automated Approach to Analyze IoT Privacy Policies . . . . . . . . . . . . . . . . . . . . . . 163 Alanoud Subahi and George Theodorakopoulos Correction to: Automated People Counting in Public Transport . . . . . . . . . . .

C1

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

The original version of this book was revised as the given name and family name of the authors in chapter 5 were corrected. The correction to the book is available at https://doi.org/10.1007/978-3030-92968-8_13

Secure Algorithm for IoT Devices Authentication Vincent Omollo Nyangaresi, Anthony J. Rodrigues, and Silvance O. Abeka

Abstract Internet of Things (IoT) security is a major concern owing to the sensitive data that flows in these networks. The fifth generation (5G) network provides high bandwidth, endearing it as an ideal underlying network for IoT communication. In addition, 5G can facilitate seamless integration of 2G, 3G, 4G, and WiFi to realize faster services, high capacity, and very short latencies. Although 5G features such as high bandwidth and seamless integration are ideal for IoT implementations, the underlying network is vulnerable to attacks such as eavesdropping, de-synchronization, sink hole, denial of service (DoS) and replay attacks, among others. To address these challenges, a number of protocols based on techniques such as elliptic curve cryptography (ECC), trusted authority, quantum cryptography, public keys, private keys, pseudonymous certificates, group handover authentication, multi-signature, and aggregate message authentication code (AMAC) technology have been proposed. Unfortunately, these protocols either have high computation and communication costs or do not provide robust security required for IoT devices communication. This renders them inefficient and susceptible to attacks such as impersonation, privacy and location sniffing, eavesdropping, session key disclosure attacks, modification, and insider attacks. Consequently, there is need for an efficient and secure key agreement and session authentication protocol for IoT deployments. In this paper, an efficient and secure handover protocol for IoT devices is proposed. The simulation results showed that this protocol exhibited lower computation and turnaround time, high stability, and moderate communications costs. It was also demonstrated to be robust against masquerading, packet replay, eavesdropping, free riding attacks, privacy and location sniffing.

V. O. Nyangaresi () Tom Mboya University College, Homabay, Kenya e-mail: [email protected] A. J. Rodrigues · S. O. Abeka Jaramogi Oginga Odinga University of Science and Technology, Bondo, Kenya © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 D. Cagáˇnová, N. Horˇnáková (eds.), Industry 4.0 Challenges in Smart Cities, EAI/Springer Innovations in Communication and Computing, https://doi.org/10.1007/978-3-030-92968-8_1

1

2

V. O. Nyangaresi et al.

Keywords Handover · IoT · Latency · Packet loss · Ping pong · Security · Authentication

1 Introduction The IoT is composed of short range communication technologies and Low-Power Wide-Area Network (LPWAN). The former consist of Bluetooth, Zigbee, WiFi, and Z-wave, while the latter is a wide-area network communication technology that can either operate on unlicensed spectrum such as LoRa and SigFox or on licensed spectrum such as 2G, 3G, and 4G cellular communication technologies [1]. According to [2], the Third Generation Partnership Project (3GPP) has devised three main IoT-related standards including Extended Coverage Global System of Mobile Communication (EC-GSM), Long-term evolution-Machine (LTE-M), and Narrow Band Internet of Things (NB-IoT). Authors in [1] explain that seamless integration of 2G, 3G, 4G, WiFi, and other access technologies can be facilitated by 5G. In addition, it supports high mobility, low latency, higher speeds (more than 10 Gigabits per second), high reliability, and increased user capacity. This renders it appropriate in IoT applications such as smart home, car networking, mobile medical, and environmental monitoring. For instance, IoT technologies use cases such as massive machine type communication (mMTC) and ultra-reliable low latency communication (URLLC) have been amalgamated in 5G to benefit from its salient features such as high bandwidth and low latencies. The 5G based IoT supports a number of novel network services including smart grid, mobile fog computing, named data networking, car-to-car communications, unmanned aerial vehicle (UAV), smart parking, and blockchain based services [3]. As pointed out in [2], scalability and long latencies are major issues in the current vehicular networks. However, through the implementation of these networks over 5G, these shortcomings are addressed. Unfortunately, ensuring security and privacy in vehicle specifics such as location and behavior is still a major drawback. Currently, most IoT systems ride on LTE/LTE-A network but efforts are being made to deploy IoT in 5G networks. Due to their close correlation with applications, dynamically changing topologies, limited resources, sophisticated network environments, and data centricity, IoT devices are vulnerable to various attacks such as free riding, impersonation, privacy sniffing, and eavesdropping attacks [1]. Although 5G is a crucial driver for IoT, little interest has been paid to its communication standard security and privacy requirements. Consequently, researchers in [4] point out that a number of issues have risen, such as the challenging task of realizing 5GIoT architecture, and enhancing security and privacy. Specifically, authentication, encryption, identity, secure storage, and secure mobility are major issues. Authors in [1] note that sensitive data transmission in IoT is prone to security attacks. It is also possible for an adversary to maliciously modify data stored in IoT devices, leading to integrity violation [5]. In addition, most IoT systems are prone to sink hole and denial of service (DoS) attacks. Moreover, ultra-densification in 5G networks using

Secure Algorithm for IoT Devices Authentication

3

small cells means that high speed IoT devices experience frequent handovers and hence high overheads [6]. Anonymity in cellular networks protects user privacy and hence if authentication protocols lack adequate level of privacy, user identity and location can be exposed [7]. For instance, during resource allocation in device discovery and source routing in device to device (D2D) communication, devices broadcast their location and identity to the network or other devices, leading to both location privacy and device privacy leak threats [1]. In 5G networks, the mobility management function (AMF) mutually authenticates the IoT’s enhanced Machine Type Communication (eMTC) devices. This authentication is via improved Extensible Authentication Protocol- Authentication and Key Agreement EAP-AKA, which facilitates secured communication between the eMTC devices and eMTC servers [2]. However, 3GPP is yet to propose how two eMTC devices can effectively establish secure communication between them [2]. As pointed out in [7], the conventional EAP-AKA’ protocol has several security issues such as simulated attacks, man-in-the-middle (MitM) attacks, and user identity privacy protection challenges. In V2X systems, during vehicle to infrastructure and vehicle to network (V2I/V2N) identity authentication, 5G-AKA protocol or EAPAKA’ protocol is utilized in ensuring mutual authentication between vehicles and the 5G core network. Unfortunately, this protocol has high bandwidth consumption, heavy signaling overhead, and slow re-authentication during handover [7]. The 3GPP plan outlines that during the first 5G phase, its new radio access network (RAN) can coexist with long-term evolution advanced (LTE-A)’s access network, sharing the Evolved Packet Core (EPC). Here, each Narrow Band Internet of Things (NB-IoT) device employs Evolved Packet System Authentication and Key Agreement (EPS-AKA) protocol for authentication through RAN and the core network. Unfortunately, LTE network’s EPS-AKA protocol has no support for IoT [8]. In addition, authors in [1] explain that the support of new technologies such as Software Defined Network/Network Function Virtualization (SDN/NFV), vehicle to everything (V2X), massive IoT, and device to device (D2D) communication introduces security issues in the 5G network. These issues include problems in 5G handover and access, lightweight security techniques, immense IoT device simultaneous security access, privacy protection, D2D’s vulnerability to privacy and security threats emanating from both cellular and ad hoc networks, need for diverse authentications for distinctive 5G-V2X settings, 5G-V2X broadcasts message security, and guaranteeing V2X to user equipment (UE) privacy. Due to direct connections among D2D proximity devices, this communication is vulnerable to both passive and active attacks since 5G’s EAP-AKA’ access authentication protocol is not ideal for such scenario [7]. Owing to frequent handovers, there is a need for more efficient mutual and handover authentication protocols among D2D UEs. Separately authenticating each IoT messages leads to high resource consumption which may exceed network capability, leading to signaling storms. In addition, the 4G network standard does not take into consideration this massive IoT device authentication. The authors in [9] explain that due to complexities in ascertaining trust across diverse network domains, fragmentation and lack of interoperability of different Low-Power Wide-

4

V. O. Nyangaresi et al.

Area Networks (LP-WANs) wireless protocols, management of handover roaming of IoT devices is cumbersome. Although 5G has promising features such as low latencies, high capacity, and bandwidths that can support V2X services, it still has numerous security and privacy challenges [10]. For instance, the conventional message authentication protocols for autonomous platoon have heavy computational and communication costs [11]. In addition, authors in [10] note that due to frequent handovers in large scale vehicular machine-to-machine (M2M) communications, security and efficient mobility management become challenging. Worse still, handovers between LTE and 802.11p have not been optimized to minimize security signaling overheads. The contributions of this paper include the following: 1. We introduce session tokens as a payload encryption constituent for confidentiality and integrity preservation. 2. GUTI is incorporated in the generation of session tokens so as to increase user anonymity. 3. We introduce nonce Rand and time stamps Tstp as context sequence information to check on the freshness of session tokens. 4. Lightweight ECC algorithms are incorporated in user authentication process and secret key exchanges for improved efficiency. 5. Exchanged payloads are encrypted using lightweight AEAD cipher to secure them against attacks. 6. We show that the developed algorithm is robust against a number of attacks and also efficient in terms of computation and turnaround time. The rest of this paper is organized as follows: Sect. 2 discusses related work, while Sect. 3 elaborates the system model employed to achieve the paper objectives. On the other hand, Sect. 4 presents results, discusses them, and evaluates the developed protocol. Lastly, Sect. 5 concludes this paper and gives future direction in this research area.

2 Related Work The deployment of EAP-AKA’ or 5G-AKA for authentication in IoT networks generates high communication and signaling overheads, and is also vulnerable to attacks. To counter these problems, researchers have suggested a number of solutions. For instance, the scheme in [12] has integrated Elliptic Curve Cryptography (ECC) and Quantum Cryptography to encrypt data transmitted across IoT devices and hence uphold availability, confidentiality, integrity and non-repudiation. However, the scheme’s feasibility and performance have not been carried out. Researchers in [13, 14] have proposed privacy aware and secure cloud assisted video reporting service to enhance road safety. This technique uses public keys, private keys, and pseudonymous certificates to guarantee vehicle authentication, traceability, and non-repudiation of sent videos, but requires Trusted Authority (TA)

Secure Algorithm for IoT Devices Authentication

5

which is centralized and hence can be a single point of failure. Its main drawback is the presence of a TA, which is assumed to be fully trusted and strongly protected so as to be difficult for any adversary to compromise it. The TA is a centralized entity, thus it potentially represents a single point of failure. To reduce signaling overhead, researchers in [15–19] have proposed group handover authentication schemes in the conventional mobile networks. These techniques build IoT device groups and then execute handover authentication in the form of groups. Unfortunately, these schemes employ asymmetric cryptography that comes with heavy computational overheads. In addition, techniques proposed in [17, 19] cannot guarantee mutual authentication hence not applicable in 5G IoT networks. Authors in [20–22] have proposed group-based access authentication schemes using group leaders that combine all access requests from group members into a single group access request and conveys it to the network. Although these schemes reduce both communication and signaling overhead, they are prone to DoS, internal forgery attacks, lack identity privacy protection and have high computational complexities due to deployment of public key cryptography (PKC). Based on multi-signature and Aggregate Message Authentication Code (AMAC) technology, authors in [23] have suggested a group authentication scheme that achieves mutual authentication between IoT devices and the network, privacy protection, and reduced signaling congestion. Researchers in [24] have used certificateless aggregation signcryption to develop data transfer and fast authentication protocol for NB-IoT system. This scheme reduced signaling overhead and assures both nonrepudiation of user identity and privacy. On the flip side, the schemes in both [23, 24] are only ideal in evolved LTE and not in 5G scenario, and the deployment of PKC result in heavy computational costs. Authors in [25] utilized acoustic waves in device discovery among IoT devices to achieve both key agreement and bidirectional initial authentication. Unfortunately, this technique lacks both identity privacy protection and strong security. In [26], a bio-inspired discovery and synchronization technique is proposed for distributed D2D discovery. However, this algorithm is inefficient, lacks key agreement and mutual authentication, and is not energy efficient. A distributed synchronization device discovery technique is proposed in [27] where proximity devices construct a synchronization group after which they publicize their presence one by one. Although the scheme shortens device discovery time, it lacks D2D security, and is prone to signaling collision. Researchers in [28] have developed a collective authentication and key agreement protocol for D2D roaming. Although it achieves privacy protection, it has heavy signaling overhead, and requires base station participation in session key derivation. Using zero knowledge proof, identity based k-anonymity secret handshake, and PKC, an authentication and key agreement protocol for group users has been developed in [29]. Although this protocol assures group anonymity and provides mutual authentication, it fails to consider roaming, and the usage of PKC results in high communication as well as computational costs. On the other hand, Hash-based Message Authentication Code (HMAC), group key agreement protocol, and pseudonym management have been utilized in [30] to

6

V. O. Nyangaresi et al.

develop a group D2D authentication technique that was shown to assure privacy. However, it results in high communication and computational costs. Authors in [31] have developed a lightweight machine-to-machine authentication protocol for industrial IoT that consist of registration and authentication and involve the router, sensor, and authentication server (AS). However, this approach is prone to both modification and insider attacks. On the other hand, researchers in [32] proposed a lightweight mutual authentication that operates in three phases, namely registration, authentication, and key establishment. The drawback of this scheme is that the registration process through which the sensor and the router register with the AS is executed over an assumed secure channel. In addition, this approach is vulnerable to session key disclosure attack. Furthermore, authors in [33] have proposed a lightweight IoT authentication scheme that was shown to be robust against session key disclosure, modification, impersonation, and eavesdropping attacks. This technique has two phases, namely registration and authentication. Unfortunately, the authors make a naïve assumption that the registration phase is secure and hence make no efforts towards securing it. In addition, the authentication server (AS) shares the devices’ real identities with other devices and hence does not uphold anonymity.

3 System Model The proposed protocol consisted of four phases, namely UE identity verification, session token generation, communication initiation, and secure payload exchanges. The implementation details of each of these phases are described in the following Sects. 3.1, 3.2, 3.3, 3.4, and 3.5.

3.1 Simulation Environment The simulated IoT key agreement and session authentication protocol involved five 5G network components including access and mobility management function (AMF), security anchor function (SEAF), user data management (UDM), general node-B (gNB), and user equipment (UE). The AMF provided UE connection and mobility management while SEAF served as an authentication intermediary between the network and the UE, validating the UE’s identity. On the other hand, UDM acted as a repository of user credentials that were employed for UE identity verification. Here, UEs are the IoT devices that communicated directly with one another while the gNB connected the UE to the 5G network as shown in Fig. 1. In scenario “1,” IoT devices are within the 5G coverage network while in scenarios “2” and “3,” the IoT devices are located outside the coverage area. The aim of these

Secure Algorithm for IoT Devices Authentication

7

Fig. 1 IoT communication scenarios

latter scenarios is to offer reliable communication services to users beyond gNB coverage area such as during emergencies and disasters. It is evident from Fig. 1 that this is a hybrid architecture that couples together both distributed and centralized access techniques. This concept renders the network susceptible to a number of security threats and attacks both from within the cellular network and the ad hoc networks. Owing to the unbound communication characteristics of IoT, the network and devices therein are exposed to air interface attacks such as masquerading, denial of service (DoS), replay, eavesdropping, and Man-in-the-Middle (MitM) attacks. To curb these attacks, this paper proposes the key agreement and session authentication protocol shown in Fig. 2 using the security features in Table 1. It is evident from Table 1 that a number of security features are incorporated in the proposed protocol to offer user anonymity, prevent eavesdropping, impersonation, location sniffing, and free riding among others.

8

V. O. Nyangaresi et al.

INPUT: SUCI, GUTI, PrKgNB, PuKgNB, PL, Rand, Tstp, UEAPpriv & UEBPpriv OUTPUT: SK, ƱKGnb, VLRes, PLencry, & PLdecry BEGIN: /* UE Identity verification */ 1. Each UE transmits ƱKgNB generation request, ƱKgNB_Req to its gNB / * 2. Upon receipt of ƱKgNB_Req, gNB forwards UE’S SUCI to the AMF/SEAF 3. AMF/SEAF contacts UDM to obtain user credentials stored in it, UDMUc 4. Using 5G-AKA, decrypt SUCI and content with UDMUc. /* User identity authentication */ 5. IF SUCIDec= UDMUc THEN: 6. Send verification results, VRRes to the gNB 7. ELSE deny ƱKgNB_Req /* Session token generation */ 8. At the gNB, generate ƱKgNB and send it to the requesting UE 9. ƱKgNB = ECDSAsign (SUCI,GUTI, PrKgNB) / * Data exchange tokens */ /* Communication initiation */ 10. Broadcast BCReq to all neighbouring devices / * Search for IoT device for data exchange */ 11. IF UE is D2D ready THEN: 12. Send BCRes to requesting UE / * Signal for data exchange availability */ 13. ELSE ignore BCReq / * ~~ TSUCIUE , TGUTIUE & TPuKgNB are the target UE’s SUCI, GUTI & gNB’s public key ~~ * / 14. Each device sends VLReq to its corresponding gNB together with its SUCI & ƱKgNB 15. gNBs exchange target UE SUCIs( TSUCIUE), GUTIs (TGUTIUE), & tokens (TƱKgNB) 16. At each gNB, validate ƱKgNB by using TSUCIUE , TGUTIUE & TPuKgNB 17. VLRes = ECDSAval (TPuKgNB, TSUCIUE, TGUTIUE) /* gNB & UE mutual authentication */ 18. IF VLRes are fine THEN: 19. Send validation results, VLRes to the requesting UEs 20. Derive secret key SK for data transmission encryption /* ~~ UEAPpriv & UEBPpriv are UEA and UEB private keys respectively ~~ */ 21. SK=(UEAPpriv, UEBPpriv) 22. Perform SK exchanges using ECDH /* Secure payload exchanges */ 23. Encrypt exchanged payload data, PL using AEAD 24. PLencry = AEADencry (SK, PL, ƱKgNB, Rand, Tstp) /* Rand, & Tstp are context sequence information */ 25. Upon receipt of PLencry, decrypt it to get the transmitted data 26. PLdecry = AEADDcry(PLencry, SK, ƱKgNB, Rand, Tstp) END

Fig. 2 Proposed protocol for secure IoT authentication

3.2 Session Keys Management IoT devices are resource constrained and as such, they require lightweight authentication algorithms. As such, elliptic curve (EC) asymmetric key algorithm which offers 128-bit cryptographic security using a 256-bit key was deployed. Although Rivest–Shamir–Adleman (RSA) is the most popular public key encryption algorithm, it was not adopted since it has 3072-bit key and hence is computational resource intensive. For digital signature generation, elliptic curve digital signature algorithm (ECDSA) is utilized as shown in step-9. ECDSA’s structure is similar to digital signature algorithm (DSA). ECDSA is a public key algorithm for generating digital signatures and whereas DSA is defined above the ring of integers, ECDSA is defined in a group of points of an EC. On the other hand, secret key exchanges were accomplished using elliptic curve Diffie–Hellman (ECDH), which is a variation of Diffie–Hellman (DH)

Secure Algorithm for IoT Devices Authentication Table 1 Proposed protocol security features

Security feature 5G-AKA ECDH SUCI GUTI TSUCIUE PuKgNB KgNB ECDSA PrKgNB SK AEAD

9 Functionality Verify the identity (SUCI) of the UE Exchange secret keys Public key encrypted IMSI for UE anonymity Temporary UE identifier for anonymity Target UE SUCI for KgNB verification gNB’s public key for KgNB verification Component of payload encryption KgNB generation KgNB generation Encrypt transmitted payload Encrypting exchanged data

algorithm for elliptic curves (ECs), as shown in step-22. Through DH cryptographic key agreement protocol, two entities with public–private key pairs on ECs can obtain a shared secret key over an unprotected communication channel. For the actual payload data encryption, lightweight authenticated encryption with associated data (AEAD) was employed as shown in phase-24. Here, the associated data included KgNB , Rand and Tstp . In the proposed protocol, each gNB had a pair of public key PuKgNB , and private keys PrKgNB . Whereas PuKgNB is shared with other gNBs in advance, PrKgNB is employed for device to device (D2D) token generation, KgNB (phase-9). To accomplish KgNB creation, ECDSA was utilized. Conventionally, 5G network employs 5G-AKA for UE identity verification and hence was utilized in this paper to authenticate the UE before KgNB generation (phase-4). The globally unique temporary identifier (GUTI) and Subscription Concealed Identifier (SUCI) were employed to offer UE anonymity (phase-9). Here, SUCI was the public key encrypted international mobile subscriber identity (IMSI) that provided the UE with encrypted permanent identifier. Both SUCI and GUTI provided the UEs with the much needed anonymity.

3.3 Session Token Generation The communication session token generation involved the IoT devices, gNB, AMF, SEAF, and UDM as shown in Fig. 3. To begin the session token (KgNB ) derivation, the UEs send a request KgNB_Req to their corresponding gNBs (step-1), which in turn forward the UE’s SUCI to the SEAF as shown in phase-2. The SEAF then utilizes the UDMUc Req to fetch user credentials UDMUc stored in the UDM (step-3). The SEAF then decrypts SUCI (phase-4) and compares the user credentials contained therein with UDMUc . Provided that user credentials are legitimate (step-5), verification results VRRes are sent to the gNB as shown in phase-6. At the gNB, KgNB is generated through ECDSA’s digital signature using

10

V. O. Nyangaresi et al.

gNB (PrKgNB)

IoT Devices (SUCI, GUTI)

ƱKgNB_Req

UDM (User credentials, UC)

SEAF UE’s SUCI

UDMUc Req

UDMUc Decrypt SUCI & compare results with UDMUc

VRRes ƱKgNB = ECDSAsign (SUCI, GUTI, PrKgNB )

ƱKgNB

Fig. 3 IoT token generation procedure Table 2 Proposed protocol exchanged messages

Message KgNB_Req VRRes BCReq BCResp VReq VRes

Functionality Request gNB to generate KgNB Inform gNB of successful SUCI authentication Probe for IoT devices for data exchange Signal availability for data exchange Request gNB to verify KgNB Informs UEs of KgNB validation

SUCI and PrKgNB after which it is forwarded to the requesting UEs (step-9). The issued KgNB adds another layer of anonymity to the UE. This ushered in the communication initiation process discussed in the following Sect. 3.4.

3.4 Communication Initiation During communication initiation and the subsequent phases, a number of messages were exchanged as shown in Table 2. The data exchange initiation process was composed of device discovery and link setup procedures. During device discovery, nearby nodes are searched using broadcast request messages, BCReq that is generated by each of them, as shown in phase-10. Upon receipt of BCReq , a node responds with BCResp (step-12), where both BCReq and BCResp contain the assigned KgNB and the UE’s encrypted identity, SUCI. In the link setup phase, peer-to-peer (P2P) connections are made between two devices.

Secure Algorithm for IoT Devices Authentication

11

Here, each device transmits a validation request, VLReq to its gNB together with SUCI and KgNB received during the discovery phase (step-14). The gNBs then exchange the target UEs SUCI (TSUCIUE ) as in phase-15 and then using each other’s pre-shared public key (TPuKgNB ), they validate KgNB using digital signatures generated by ECDSA without connecting to the core network (step-17). This is followed by the transmission of validation results VLRes to the requesting UEs as in phase-19. After validation, secret key SK is generated using the UEs private keys (UEA Ppriv and UEB Ppriv ) as in step-21 and then ECDH is employed to securely exchange this SK to safeguard data exchanges (phase-22). As such, even if an adversary eavesdrops data transmissions at the middle of the key exchange, SK cannot be derived.

3.5 Secure Payload Exchanges During the actual data transfers, security is ensured by encrypting data using AEAD cipher prior to transmission as in phase-24. For the encryption, UES utilize their KgNB , SK, plain text payload (PL), and context sequence information (CSI) to uphold data confidentiality and integrity. Here, CSI consisted of random parameter, Rand (nonce), and time stamps, Tstp which checked on the freshness of KgNB to prevent replay attacks. An IoT device receiving PLencry has to decrypt it to retrieve the transmitted message as shown in step-26. Figure 4 shows the proposed secure IoT data exchange process. As shown in Fig. 4, lightweight AEAD cipher is employed to encrypt the payload PL to offer confidentiality, authenticity and integrity in the communication process. This creates a message authentication code (MAC) for checking data integrity. As shown here, AEAD uses communication session specific data and also information of the other party and hence provided authenticity that assures that PL is being sent from the legitimate party at the right moment. In this paper, additional session specific information, timestamp Tstp and random parameters, Rand were introduced to curb replay attacks. Consequently, the resulting data format is as shown in Fig. 5. Here, header represents the transmission control protocol (TCP) header which contains fields such as source port, destination port, sequence number, acknowledgment number, header length, control flags (URG, ACK, PSH, RST, SYN, and FIN), window size, checksum, options and padding, and urgent pointer.

4 Results and Discussion In this section, the obtained simulation results are presented and discussed. In Sect. 4.1, the comparison of the lightweight ECC implemented in this paper is compared with both DSA and RSA in terms of time complexity. These two are popular

12

V. O. Nyangaresi et al.

SgNB [PuSKgNB]

TgNB [PuTKgNB]

UEB [SUCIB, GUTIB]

UEA [SUCIA, GUTIA]

BCReq ( ƱSKgNB, SUCIA)

BCRes (ƱTKgNB, SUCIB)

VReq, ƱSKgNB, SUCIA

VReq( ƱTKgNB, SUCIB)

VLRes = ECDSA (SPuKgNB, SUCIA, GUTIA)

VLRes = ECDSA (TPuKgNB, SUCIB, GUTIB)

VLRes

VLRes Exchange SK using ECDH

Transmit PL encrypted by AEAD, PLencry

Decrypt PLencry

Fig. 4 Proposed secure payload exchange process

Header

PLencry

AEAD’s MAC

ƱKgNB

Tstp & Rand

Fig. 5 Exchanged payload format

algorithms for digital signature generation and hence it was necessary to discern how they perform in comparison with ECC algorithms employed in this paper. On the other hand, Sects. 4.2, 4.3, 4.4, 4.5, and 4.6 present and discuss computation time, turnaround time, stability, communication overheads, and security evaluation of the developed protocol, respectively.

4.1 Comparison with DSA and RSA In this section, the lightweight ECC algorithm which was employed in the developed protocol is compared with other most deployed digital signature algorithms, namely DSA and RSA. The time complexity results obtained for different packet sizes are

Secure Algorithm for IoT Devices Authentication

13

Fig. 6 Time complexities comparisons

shown in Fig. 6. Based on the results in Fig. 6, it is evident that different packet sizes yielded varied time complexities for all the three algorithms. Generally, as the packet sizes increased, there was a corresponding increase in the algorithm time complexities. Among these algorithms, DSA had the largest time complexity followed by RSA algorithm and lastly the ECC algorithm. For small packet sizes, ECC and RSA exhibit almost similar time complexities. However, as the packet sizes increase, the time complexities of the two algorithms become more pronounced.

4.2 Computation Time This represented the time needed by the IoT entities to calculate all the exchanged messages during the authentication process at their ends. To accomplish this, the message execution time was observed in the local host running Windows 7 operating system on Core i3 processor, which were the same operating systems and core configurations that the protocols developed in [31–33] were executed. Thereafter, comparisons were made against computation time for other lightweight IoT algorithms developed in [31–33] as shown in Table 3 and Fig. 7. As shown in Table 3, the protocol developed in [32] had the highest computation time followed by the protocol proposed in [33]. The approach in [32] employs advanced encryption standard (AES) encryption algorithm hence the high computational time. On the other hand, the proposed protocol had the least computation time

14 Table 3 Computation time comparisons

V. O. Nyangaresi et al. Protocol [31] [32] [33] Proposed protocol

Computation time (ms) 170 2380 185 168

Fig. 7 Computation time comparisons

of 168 ms, which was fairly close to the value obtained with the protocol developed in [31]. Although the scheme in [33] has higher computation time than the scheme in [31], it offers protection against privacy sniffing, impersonation, eavesdropping, location sniffing, and free riding attacks. As such, the protocol in [33] can protect against packet modification and session hijack attacks, a feature that lacks in the scheme proposed by [31].

4.3 Turnaround Time This refers to the total duration taken by the protocol entities to execute the authentication process. To achieve this, the authentication duration for the UE with SEAF, UE with gNB and mutual authentication between gNBs was measured. The results were then compared with values obtained for protocols developed in [31–33] as shown in Table 4 and Fig. 8. Based on the results in Table 4, the mean turnaround time was highest for the scheme developed in [32] due to the deployment of AES message encryptions, followed by the protocol proposed in [33]. The proposed protocol had the least turnaround time of 1387 ms. The protocol in [33] took 19 ms more than that in [31] since there are three parties involved during the authentication process as compared to two entities (sensor and router) in [31].

Secure Algorithm for IoT Devices Authentication Table 4 Turnaround time comparisons

15 Protocol [31] [32] [33] Proposed protocol

Turnaround time (ms) 1441 3855 1460 1387

Fig. 8 Turnaround time comparisons

Although the approach in [31] has relatively better turnaround time, it is susceptible to modification and session key disclosure attacks, something that is addressed by the protocol in [33]. Further evaluation was executed to determine how each protocol’s performance was affected by varying packet sizes. The results are shown in Fig. 9. It is evident, based on the graphs in Fig. 9, that all protocols consumed varying turnaround times as the packet sizes were slowly increased from 50 bytes to 200. Among all these protocols, the one proposed in [32] had the most pronounced turnaround time increment as the packet sizes were increased. This can be attributed to increasing AES encryption of the now many packets. On the other hand, although the turnaround time for other protocols also increased with packet sizes, the increments were not so much pronounced.

4.4 Stability To access how each of the four protocols’ performance varied under different conditions in terms of turnaround time, standard deviation (S) was utilized to compute stability of each protocol for different packet sizes. Taking n as the number

16

V. O. Nyangaresi et al.

Fig. 9 Turnaround comparisons for varying packet sizes

of iterations, S was computed as shown in (1). 



S=

(x − x)2 n−1

(1)

The idea was to execute the protocols 20 times (n = 20) using a variety of packet sizes as the variance of the turnaround was recorded for each protocol. The results obtained are shown in Fig. 10. In stability evaluations, a large value for standard deviation implied fluctuations in turnaround and hence instability of the performance of the underlying protocol. Based on the results of Fig. 10, the scheme developed in [32] had the highest S value, and hence was the most unstable followed by the protocol proposed in [33]. On the other hand, our protocol had the least S value and therefore was the most stable.

4.5 Communication Overheads The communication overheads denoted the sum of all bits utilized for exchanging messages among all protocol parties. In the proposed protocol, messages were exchanged among UEA , UEB , source gNB (SgNB), target gNB (TgNB), and SEAF as shown in Table 5, which also presents a breakdown of the message exchanges in the proposed protocol. Here, the bandwidth needed for executing communication in

Secure Algorithm for IoT Devices Authentication

17

Fig. 10 Turnaround stability comparisons Table 5 Communication overheads comparisons Communication path Messages Token generation UE → gNB KgNB_Req gNB → SEAF UE’s SUCI SEAF → UDM UDMUc Req UDM → SEAF UDMUc SEAF → gNB VRRes gNB → UE KgNB Communication initiation and payload transmission UEA → UEB SKgNB , SUCIA UEB → UEA TKgNB , SUCIB UEA → SgNB SKgNB , SUCIA UEB → TgNB TKgNB , SUCIB SgNB → UEA VLRes TgNB → UEB VLRes UEA ↔ UEB SK UEA → UEB PLencry Total

Communication overhead 1 × 128 = 128 1 × 128 = 128 1 × 128 = 128 1 × 128 = 128 1 × 128 = 128 1 × 128 = 128 2 × 128 = 256 2 × 128 = 256 2 × 128 = 256 2 × 128 = 256 1 × 128 = 128 1 × 128 = 128 1 × 128 = 128 1 × 128 = 128 2304 bits

the proposed protocol is the sum of all the bits exchanged during key agreement, session token generation, communication initiation, and payload exchanges. All messages exchanged were ciphered using the keys shown in Table 1 and each cipher was 128 bits long. The length of each cipher was maintained at 128 bits so that the session key generated at the end of the authentication process is 128

18 Table 6 Communication overheads

V. O. Nyangaresi et al. Protocol [31] [32] [33] Proposed protocol

Communication overheads (bits) 1920 2048 2048 2304

Fig. 11 Communication overheads comparisons

bit long. Based on the computations in Table 5, the total communication overhead was 2304 bits. The communication overheads of the developed protocol were then compared to schemes developed in [31–33] as shown in Table 6 and Fig. 11. Based on these results, the protocol developed in [31] had the least communication overhead of 1920 bits followed by protocols proposed in [32, 33] which had similar communication overheads of 2048 bits. On the other hand, the proposed protocol had the highest communication overheads. However, this high bandwidth cost is justified since the proposed protocol executes mutual authentication among all the IoT entities and also encrypts the exchanged payload hence is robust against replay attacks, location sniffing, free riding, eavesdropping, privacy sniffing, impersonation attacks among others. As such, the developed protocol is more secure compared with schemes developed in [31–33]. For instance, in [31], only two parties are authenticated using 1920 bits while our protocol authenticates five entities (two UEs, SgNB, TgNB, and SEAF) using 2304 bits.

Secure Algorithm for IoT Devices Authentication

19

4.6 Security Evaluation The proposed protocol executed mutual authentication among all the communication entities as shown in Fig. 2 in Sect. 3.2. To evaluate the security of the proposed protocol, six threat and attack models were used: location and privacy sniffing, masquerade, free riding, eavesdropping, and packet replay. In location sniffing, an adversary broadcasts BCReq with invalid location information with the goal of disrupting the IoT communication process during the device discovery phase. To curb this, BCReq has to be verified before any response is given by other IoT devices. In the developed protocol, all devices are authenticated using SUCI before KgNB tokens are generated by the gNB, as shown in Fig. 2. As such, each UE is authenticated by its corresponding gNB. In addition, the generated KgNB tokens are also validated using VReq and VRes . Moreover, the transmitted payload is ciphered using AEAD and hence gNBs can easily deny access to malicious UEs. In packet sniffing, an attacker uses BCReq to search and track IoT devices. To thwart this attack, all devices must be authenticated and they also need to have anonymous identity. In the developed protocol, KgNB tokens are generated based on UE’s SUCI, GUTI, PrKgNB (gNB’s digital signature) using ECDSA. As such, SUCI, GUTI, and KgNB ensure UE anonymity and hence an adversary is unable to discern the true UE identity. Regarding masquerade attacks, a malicious UE pretends to be a legitimate UE by forging other UE’s identity such as the IMSI. To curb this attack, strong authentication of all UE’s should be implemented. In the developed protocol, each UE is issued with KgNB signed by its corresponding gNB using ECDSA algorithm. Before KgNB generation, the identities of all UEs are verified by decrypting their SUCI (SUCIDec ) and comparing the resulting values with the ones stored in the UDM (UDMUc ). Upon completion of this authentication, gNBs generate KgNB tokens using their private keys (PrKgNB ). Consequently, it becomes cumbersome for an adversary to masquerade as a validated UE. In free riding attacks, upon receipt of required data from other UEs, selfish UEs fail to share their resources due to energy consumption. This compromises system availability and to mitigate this attack, UEs identities must be verified and controlled by the gNBs. In the proposed protocol, all UEs are verified using their SUCI and the validity of KgNB is also checked by the gNBs. The goal of eavesdropping is to passively listen to the message exchanges among the UEs. On the other hand, in replay attacks, an adversary can capture the transmitted data, fabricate it before forwarding it to other UEs. These two attacks can be prevented by boosting confidentiality and integrity through proper encryption of the exchanged messages. This was achieved in the developed protocol using AEAD cipher to encrypt the payload. During this process, SK, PL, KgNB , Rand , and Tstp are used. Here, Rand and Tstp represent a nonce and time stamp, respectively, that are used as CSI to check on the freshness on the generated KgNB . These two parameters together with the generated KgNB form associated data that was employed to generate MAC used by the UEs to verify integrity of the exchanged

20

V. O. Nyangaresi et al.

messages as well as the validity of the sender UEs. This effectively thwarts both eavesdropping and packet replay attacks. On the other hand, the protocol in [31] is susceptible to insider and modification attacks while the scheme in [32] is vulnerable to session key disclosure attack. Modification and session key disclosure attacks in [31] facilitate adversarial mapping of the network since authentication request responses always emanate from the router. In addition, this protocol is unable to activate authentication process for second hop parties in the network. However, in the developed protocol, any two devices can run the authentication protocol. The protocol in [31] made an assumption that the sensor devices are legitimate. Since this is not always the case due to existence of masquerading UEs with an IoT environment, it is feasible for any sensor to masquerade as a router through the extraction of router’s secret key. However, in the developed protocol, IoT devices are unable to extract secret parameters belonging to other devices within the network. This is because secret parameters are only known to the gNB and UDM 5G components. On the other hand, the protocol in [33] makes a naïve assumption that the registration phase is secure and hence makes no efforts towards securing it. Furthermore, the AS shares the devices’ real identities with other devices in [33] and hence do not provide UE anonymity.

5 Conclusion and Future Work The goal of this paper was to design, develop, and simulate a secure key agreement and session authentication protocol for IoT devices. This was an effort directed towards solving security issues surrounding IoT’s device discovery, link setup, and payload exchanges phases. These security challenges include lack of authentication for verifying UEs identity, and lack of payload encryption. This compromises both confidentiality and integrity, rendering the network susceptible to privacy sniffing, eavesdropping, masquerade, location spoofing, and free riding attacks. The amalgamation of IoT technologies use cases such as mMTC and URLLC in 5G make the security challenges above more critical and cumbersome to counter. This is because conventional security techniques cannot be deployed or processed optimally owing to resource constrained nature of IoT devices in form of power consumption, memory, and performance. This is worsened by the fact that IoT applications handle highly sensitive and private data. To overcome these security and performance issues, a need arose for a lightweight and proper authentication protocol among the UEs to uphold confidentiality, anonymity, and integrity. As such, ECC which is the most popular lightweight asymmetric algorithm was deployed in the developed protocol. The simulation results showed that our protocol was robust against these attacks. In addition, the performance evaluation results showed that the ECC deployed here had the least time complexity in comparison with DSA and RSA. The developed protocol also had the least computation costs, turnaround times, and stability, with moderate communication overheads. Future work in this area involves deploying the developed protocol in a real IoT communication

Secure Algorithm for IoT Devices Authentication

21

environment to facilitate the validation of the simulation results obtained. There is also need for the evaluation of the developed protocol using other parameters such as energy consumptions as well as other attack models.

References 1. Cao J, Ma M, Li H, Ma R, Sun Y, Yu P, Xiong L (2019) A survey on security aspects for 3GPP 5G networks. IEEE Commun Surv Tutorials 22(1):170–195 2. Sicari S, Rizzardi A, Coen-Porisini A (2020) 5G in the Internet of Things era: an overview on security and privacy challenges. Comput Netw 2020:107345 3. Ferrag M, Maglaras L, Argyriou A, Kosmanos D, Janicke H (2018) Security for 4G and 5G cellular networks: a survey of existing authentication and privacy-preserving schemes. J Netw Comput Appl 101:55–82 4. Li S, Da Xu L, Zhao S (2018) 5g internet of things: a survey. J Ind Inf Integr 10:1–9 5. Khan MA, Salah K (2018) IoT security: review, blockchain solutions, and open challenges. Futur Gener Comput Syst 82:395–411 6. Boujelben M, Rejeb S, Tabbane S (2015) A novel green handover self-optimization algorithm for LTE-A/5G HetNets. In: 2015 international wireless communications and mobile computing conference (IWCMC). IEEE, Piscataway, pp 413–418 7. Cao J, Ma M, Li H, Zhang Y, Luo Z (2014) A survey on security aspects for LTE and LTE-A networks. IEEE Commun Surv Tutorials 16(1):283–302 8. Saxena N, Grijalva S, Chaudhari NS (2016) Authentication protocol for an IoT-enabled LTE network. ACM Trans Internet Technol 16(4):1–20 9. Torroglosa-Garcia EM, Calero JMA, Bernabe JB, Skarmeta A (2020) Enabling roaming across heterogeneous IoT wireless networks: LoRaWAN meets 5G. IEEE Access 2020:17 10. Lai C, Lu R, Zheng D, Shen XS (2020) Security and privacy challenges in 5G-enabled vehicular networks. IEEE Netw 34(2):37–45 11. Jo HJ, Kim IS, Lee DH (2018) Reliable cooperative authentication for vehicular networks. IEEE Trans Intell Transp Syst 19(4):1065–1079 12. Khan A, Abdullah J, Khan N, Julahi A, Tarmizi S (2017) Quantum-elliptic curve cryptography for multihop communication in 5g networks. IJCSNS 17(5):357–365 13. Eiza MH, Ni Q, Shi Q (2016) Secure and privacy-aware cloud-assisted video reporting service in 5g-enabled vehicular networks. IEEE Trans Veh Technol 65(10):7868–7881 14. Mohseni-Ejiyeh A, Ashouri-Talouki M (2017) Sevr+: Secure and privacy-aware cloudassisted video reporting service for 5g vehicular networks. In: Iranian conference on electrical engineering (ICEE). IEEE, Piscataway, pp 2159–2164 15. Cao J, Li H, Ma M, Li F (2018) UPPGHA: uniform privacy preservation group handover authentication mechanism for mMTC in LTE-A networks. Secur Commun Netw 2018:1–16 16. Cao J, Li H, Ma M, Li H (2017) G2RHA: group-to-route handover authentication scheme for mobile relays in LTE-A high-speed rail networks. IEEE Trans Veh Technol 66(11):9689–9701 17. Cao J, Li H, Ma M (2015) GAHAP: A group-based anonymity handover authentication protocol for MTC in LTE-A networks. In: 2015 IEEE international conference on communications (ICC). IEEE, Piscataway, pp 3020–3025 18. Cao J, Li H, Ma M, Li F (2015) UGHA: uniform group based handover authentication for MTC within E-UTRAN in LTE-A networks. In: 2015 IEEE international conference on communications (ICC). IEEE, Piscataway, pp 7246–7251 19. Kong Q, Lu R, Chen S, Zhu H (2017) Achieve secure handover session key management via mobile relay in LTE-advanced networks. IEEE Internet Things J 4(1):29–39 20. Cao J, Ma M, Li H (2015) GBAAM: group-based access authentication for MTC in LTE networks. Secur Commun Netw 8(17):3282–3299

22

V. O. Nyangaresi et al.

21. Li J, Wen M, Zhang T (2016) Group-based authentication and key agreement with dynamic policy updating for MTC in LTE-A networks. IEEE Internet Things J 3(3):408–417 22. Lai C, Li H, Lu R, Jiang R, Shen X (2013) LGTH: a lightweight group authentication protocol for machine-type communication in LTE networks. In: 2013 IEEE global communications conference (GLOBECOM). IEEE, Piscataway, pp 832–837 23. Cao J, Ma M, Li H, Fu Y, Liu X (2018) EGHR: efficient group-based handover authentication protocols for mMTC in 5G wireless networks. J Netw Comput Appl 102:1–16 24. Cao J, Yu P, Ma M, Gao W (2018) Fast authentication and data transfer scheme for massive NB-IoT devices in 3GPP 5G network. IEEE Internet Things J 6:1561–1575 25. Xie P, Feng J, Cao Z, Wang J (2018) Genewave: fast authentication and key agreement on commodity mobile devices. IEEE/ACM Trans Networking 26(4):1688–1700 26. Chao SL, Lee HY, Chou CC, Wei HY (2013) Bio-inspired proximity discovery and synchronization for D2D communications. IEEE Commun Lett 17(12):2300–2303 27. Huang PK, Qi E, Park M, Stephens A (2013) Energy efficient and scalable device-to-device discovery protocol with fast discovery. In: 2013 IEEE international workshop of Internet-ofThings networking and control (IoT-NC). IEEE, Piscataway, pp 1–9 28. Wang M, Yan Z, Niemi V (2017) UAKA-D2D: universal authentication and key agreement protocol in D2D communications. Mobile Netw Appl 22(3):510–525 29. Hsu RH, Lee J, Quek TQ, Chen JC (2018) GRAAD: group anonymous and accountable D2D communication in mobile networks. IEEE Trans Inf For Secur 13(2):449–464 30. Wang M, Yan Z (2018) Privacy-preserving authentication and key agreement protocols for D2D group communications. IEEE Trans Ind Inf 14(8):3637–3647 31. Esfahani A, Mantas G, Matischek R (2017) A lightweight authentication mechanism for M2M communications in industrial IoT environment. IEEE Internet Things J 6:288–296 32. Khemissa H, Tandjaoui D (2016) A novel lightweight authentication scheme for heterogeneous wireless sensor networks in the context of Internet of Things. In: Wireless telecommunications symposium (WTS). IEEE, Piscataway 33. Adil A, Mazhar A, Abdul NK, Tauqeer K, Faisal R, Yaser J, Junaid S (2019) A multi-attack resilient lightweight IoT authentication scheme. Trans Emerg Telecommun Technol 2019:1–15

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route in Multimodal Transport Systems Olexiy Pavlenko , Dmitriy Muzylyov , Natalya Shramenko ˇ Dagmar Cagánová , and Vitalii Ivanov

,

Abstract The paper aims to carry out the mathematical modeling of the goods delivery process from China to Ukraine, using possible organization options of such work and advantages of existing routes using different kinds of transport. The paper presented successful mathematical modeling of goods delivery from China to Ukraine to determine effective options. The structure of goods delivery from China to Ukraine has been designed in the form of four alternative routes (options), which considering the use of railway, maritime, road and air transport, and related infrastructure (stations, ports, warehouses, terminals, customs). It has been found that values of order delivery volumes of the corresponding type of goods based on parameter analysis of orders flow for trade enterprises of Kharkiv, cargo transportation volume in the current batch, risk assessment factor using related kinds of transport, and goods delivery time for each option. As an experiment result, enterprises’ profit was obtained using initial and final values of the unit of the good according to proposed options. It was taken into account in regression model designing, which allowed determining the best route of transportation. Keywords Option · Logistics · Goods · Data · Transport · Delivery · Simulation · Information · Regression analysis · Model O. Pavlenko Kharkiv National Automobile and Highway University, Kharkiv, Ukraine D. Muzylyov Kharkiv Petro Vasylenko National Technical University of Agriculture, Kharkiv, Ukraine N. Shramenko Kharkiv Petro Vasylenko National Technical University of Agriculture, Kharkiv, Ukraine Ukrainian State University of Railway Transport, Kharkiv, Ukraine D. Cagáˇnová () Slovak University of Technology in Bratislava, Slovakia e-mail: [email protected] V. Ivanov Sumy State University, Sumy, Ukraine e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 D. Cagáˇnová, N. Horˇnáková (eds.), Industry 4.0 Challenges in Smart Cities, EAI/Springer Innovations in Communication and Computing, https://doi.org/10.1007/978-3-030-92968-8_2

23

24

O. Pavlenko et al.

1 Introduction The National Transport Strategy of Ukraine identifies priority areas for improving the quality of transport services. It calls for their provision and infrastructure development to be brought closer to European standards: increased level of safety and reduction of negative impact on the environment [1]; response to the need for improved governance [2]; administrative reform and decentralization of the tasks and functions of the central executive branch; implementation of anti-corruption policy, corporate governance in the public economy sector [3]. It will improve the service quality of orders for goods delivery in international traffic, including from China to Ukraine. The geographical structure of foreign trade in goods in January–July 2019, according to the State statistics (Table 1), shows that China occupies third place in terms of goods exports, and duration of imports, take the first place [4]. However, this country’s largest increase in foreign trade is also demonstrated— exports are 160%, and imports are 127%. Therefore, it is necessary to give more attention to this direction of delivery of goods, taking advantage of all kinds of transport (road, railway, water, and air), and considering the technical capabilities of countries involved in delivery processes: ports (departure, destination, transit), railway connection stations, customs, terminal complexes, etc.). The necessity to design a new system of materials management is based on the application of world scientific experience in using SMART solutions [5] and on the implementation of integration processes of production [6], transportation [7], storage, and sale of export–import cargo flows [8]. According to logistics operators’ opinion, demand is growing today in trading enterprises for supplies of goods from China (PRC). It should be noted that in many developed countries of the world, the crisis in previous years has had a positive impact on cost savings in production processes and has encouraged producers to implement a just-in-time model [9]. Goods supply is now relevant for the international communication market in a shorter time and specific hours, and with different batches, which obviously cannot happen without the consolidation of deliveries by other customers from various suppliers using some kinds of transport. Therefore, there is a need to design new scientifically grounded approaches to determine an effective option for goods delivery from China to Ukraine.

2 Literature Review All Ukraine regions need constant supplies from abroad of various resources and goods requiring the delivery of equipment, materials, industrial goods, and other resources in small batches, as their consumption is limited and accumulation and storage are inefficient. Transportation of piece packaged cargoes takes an essential place in the transport services of economies globally. These cargoes provide all

Country 1. Poland 2. Russian Federation 3. China 4. Turkey 5. Italy

Export 1000 × USA dollars 1,976,010.8 1,889,939.6 1,868,468.2 1,530,199.1 1,487,022.1 In % to Jan–Jul 2018 104.2 88.0 160.4 97.7 90.1

Import 1000 × USA dollars 2,255,956.1 4,434,110.5 4,803,915.5 1,071,820.4 1,136,602.7 In % to Jan–Jul 2018 110.7 99.9 127.5 126.8 104.4

Balance, 1000 × USA dollars −279,945.3 −2,544,170.9 −2,935,447.3 458,378.7 350,419.4

Table 1 Geographical structure of foreign trade in goods in January–July 2019 by exports volume (an example of five countries)

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route. . . 25

26

O. Pavlenko et al.

economic branches with the necessary resources, raw materials, and materials [10– 12]. The main advantages of prefabricated transportation are the profitability and convenience of transport in this format. Significant savings were achieved through the efficient use of valuable cargo loading space. In this case, even air travel, which is generally high in price, maybe quite affordable. Many companies in Ukraine do not have their extensive network in the country, but they are engaged in supplying goods from abroad, including China. Simultaneously, they are used motor, railway, marine, and air transport [13–15]. The schemes are built considering possible delivery options at minimum cost and topics, considering the ports operation’s specifics in China, Ukraine, railway messages, and storage complexes. Multimodal transport is a key component of modern logistics systems, especially for international long-distance transport. Document [16] discusses various alternative routes for the export of laptops from Chongqing (China) to Rotterdam (Netherlands). He chooses seven available routes to deliver laptops from Chongqing to Rotterdam. The multimodal model was adopted to demonstrate alternative routes using various transportation costs, freight cost, travel time, transportation distance, document fees, charges for transshipment in ports, customs duties, confidence index, etc. The theoretical-game approach is used [17, 18] to solve problems to optimize China’s delivery of goods. Simulation of logistics systems using neural networks and Blockchain [19–22] and Petri networks [23–26], as well as the necessary condition, is using mathematical modeling of transport processes [27–29] and special software, for example, FlexSim [30], and MATLAB [31]. The choice of supply chain options is proposed by some researchers based on technological aspects of transportation [32–34] and considering technical nuances of rolling stock using [35]. In particular, we have used the example of railway transport to intellectualize its traffic management [36]. It is also necessary to consider the influence of external environment factors on delivery processes [37]. Data Science methods [38] give effective solutions to business problems, which are the same as logistics tasks. Simultaneously, the principles must use greater homogenization of information technology and software for small transport companies, same as for various small manufacturing enterprises how this conception is described in research [39]. It gives a possibility to find a flexible management decision to choose a rational supply scheme. Many scientific papers propose a system of product supply chain tracking, which has many connections and complex factors while optimizing models of studied objects using modern information systems and technologies [40–43].

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route. . .

27

3 Research Methodology The paper aims to carry out the mathematical modeling of the goods delivery process from China to Ukraine, using possible organization options of such work and advantages of existing routes using different kinds of transport. The mathematical modeling method allows us to design new and improve technological processes for operation in optimal modes. Such an approach reduces the study of the phenomena of the outside world to math problems. Fundamental changes are required concerning users of services using different transport kinds to increase the transport branch’s efficiency. Implementation of modern transport process technologies is aimed at the high quality of transport services, which can be achieved by activating several kinds of transport. It is possible to reduce cargo delivery and cost-minimizing [44–46]. China has become one of Ukraine’s main trading partners. The People’s Republic of China (PRC) has an active subsidy policy within the framework of the “Made in China 2025” project, which predicts the country to enter the absolute leaders of world production (more than 40% of all goods). Government subsidies that stimulate exports come from primary industrial goods imported into Ukraine [47]. The value share of Chinese goods in Ukrainian imports increased from 11% to 34% (to 27 billion US dollars) in 2018, beating both Europe and Russia according to this indicator. Imports volume from China is also growing much faster than Ukrainian exports in this country [48]. The commodity structure of imports from China (Fig. 1) shows a significant volume of supplies of such goods group “Machines, equipment, and mechanisms; electrotechnical equipment; their components”—29.1% of the total volume. Diodes, transistors, photosensitive semiconductor devices, and piezoelectric crystals have the largest presence from previous goods group in the Ukrainian market of Chinese. Here the volume of deliveries already reaches 2 billion UAH per month, which is more than 90% of all Ukrainian imports of related products [49]. Activity analysis of most buyers’ companies of goods from China in Kharkiv showed that “Machines, equipment, and mechanisms; electrotechnical equipment; their components” is the largest set of orders for 85 commodity groups of commodity nomenclature for foreign economic activity. “Motors and generators electric” are the largest number of orders for the subgroup in this group, up to 20% (Fig. 2). The cargo recipients are located in Ukraine in Kharkiv; consignors are in China— it is a departure warehouse (consolidation) in Guangzhou. The research object (a delivery process of goods from China to Ukraine) is carried out based on existing networks of roads, railway stations, ports, and the Silk Road’s existence [50]. Figure 3 shows the following routes (variants): 1. “Route 1”: Guangzhou, China (railway station)—Chongqing, China (railway station)—Warsaw, Poland (railway station)—terminal Kharkiv, Ukraine; 2. “Route 2”: Guangzhou, China (seaport)—Odessa, Ukraine (seaport)—terminal Kharkov, Ukraine;

28

O. Pavlenko et al.

Fig. 1 Share of total goods imports from China to Ukraine

Fig. 2 Share of goods on orders at enterprises of Kharkiv

3. “Route 3”: Guangzhou, China (seaport)—Koper, Slovenia (seaport)—terminal Kharkiv, Ukraine; 4. “Route 4”: Guangzhou, China (airport)—Boryspil, Ukraine (airport)—terminal Kharkiv, Ukraine. It is proposed to use profit (P) of selling goods as an estimated indicator that are delivered from China to Ukraine by choosing an effective option (route):   P = max Zi − Vij − Si , i = 1 . . . n; j = 1 . . . m, j

(1)

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route. . .

29

Fig. 3 Routes of goods delivery from China to Ukraine

where Zi final cost of ith good, UAH/kg, n number of specific goods group, Vij delivery expenses on ith good transported by jth route (option) (UAH/kg), m number of the specific route, Si initial cost of ith good (UAH/kg). The initial and final cost is determined by the customer per unit of goods—in our case, we accept per kilogram—this is important in determining the final rates when chartering containers, trucks, sea ships or river vessels, and airplanes. The delivery cost on ith good transported by jth route (option): K 

Vij =

k=1

k + VDj

C  c=1

VFc Rj + Vaddj

qi

,

(2)

k total costs for performance of kth operations on delivery of ith goods where VDj along a specific route (UAH), K number of operations on the corresponding route (variant), VFc Rj total expenses for container chartering, load unit (pallet, box), as well as reservation of space in trucks, ships, railway carriage (UAH), C type of lading according to corresponding transport, means, Vaddj costs for additional operations on the corresponding route (UAH), qi delivery order quantity of the relevant item type (kg). Cost components consider the cost of performance of the corresponding types of works, the time of their performance, risk components, operation scopes (distances of transportation, times of loading-unloading, storage, etc.) on all logistics sections chain of the goods delivery.

30

O. Pavlenko et al.

4 Results and Discussion The methodology was chosen to determine the effective delivery option. It consists of the following steps: designing alternatives for goods delivery, determining an evaluation parameter for selecting an effective choice; building models of options for goods delivery; modeling delivery process, and choosing an estimate; selecting an effective option based on specific criteria. The natural observation was carried out during studies of orders for trading enterprises of Kharkiv. The analysis allowed to determine such parameters of order flow: quantity of delivery order of corresponding goods type (qi ), an amount of cargo transportation in a specific lot (Q), risk assessment factor with the participation of related kinds of transport (Rjt ), time of goods delivery for each option (Td ). The required number of observations has been determined (Table 2) to obtain the most reliable data on change process parameters values. It has been established based on the analysis of the order flow parameters that value of delivery order volume of the corresponding goods type, cargo volume transportation in current batches, risk assessment factor with the participation of related kinds of transport, and goods delivery time for each variant which are distributed according to the ordinary law of distribution by random values. It was confirmed by statistical analysis of the data carried out using software Statistica. Four factors were identified using the analytical models built that affect delivery efficiency, as well as minimum and maximum values were determined: order quantity for delivery of the corresponding goods type, cargo transportation volume in current batches, risk assessment factor with the participation of related kinds of transport, goods delivery time for each option (Table 3). It was designed a corresponding plan of obtained data (Table 4) based on previous results (Table 3) and simulation principles of full-factor plans for experiments. The regression model was determined in linear form based on results analysis of the experiment, in which model each coefficient indicates influence levels of

Table 2 Results of sample size calculations

Indicators Mathematical expected value Root-meansquare deviation Error of calculations Sample size (unit)

Delivery order quantity of the relevant item type (kg) 70.37

Transportation volume in the current batch (kg) 6420

67.70

4377.64

2.76

394.66

3.52

321.00

0.06

43.20

60

52

The modal risk assessment factor 1.255

39

Delivery time (h) 864

62

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route. . .

31

Table 3 Influence factor variation levels

Level of values Maximal Minimal

Delivery order quantity of the relevant item type (kg) 140 1.74

Transportation volume in the current batch (kg) 1258 260

The modal risk assessment factor 1.5 1.01

Delivery time (h) 1392 336

The modal risk assessment factor 1.01 1.01 1.5 1.5 1.5 1.01 1.01 1.5 1.01 1.5 1.01 1.01 1.5 1.5 1.01 1.5

Delivery time (h) 336 1392 1392 1392 1392 336 1392 1392 1392 336 336 336 336 336 1392 336

Table 4 Influence factor variation levels (real values)

Observation series Series 1 Series 2 Series 3 Series 4 Series 5 Series 6 Series 7 Series 8 Series 9 Series 10 Series 11 Series 12 Series 13 Series 14 Series 15 Series 16

Variation levels Delivery order quantity of the relevant item type (kg) 1.74 1.74 1.74 1.74 140 140 140 140 140 140 1.74 140 1.74 140 1.74 1.74

Table 5 Regression models for four delivery options

Transportation volume in the current batch (kg) 260 260 260 1258 1258 260 260 260 1258 260 1258 1258 260 1258 1258 1258 Option “Route 1” “Route 2” “Route 3” “Route 4”

Regression model P = 2711 + 19.81 · q − 1.33 · Q − 4.16 · Td P = 2632 + 22.8 · q − 1.55 · Q − 4.17 · Td P = 2607 + 19.92 · q − 1.27 · Q − 4.16 · Td P = 2286 + 78.8 · q − 6.56 · Q − 4.17 · Td

the corresponding factor on effective indicators (Table 5). This model is the most adequate because the R-square value for the first three variants is near 1, and the fourth one is 0.795. The values of regression model coefficients were also checked by values of standard error, “t-statistics,” “P-value,” “minimal and maximal values”—the indicator of risk assessment with the participation of corresponding kinds of transport is not considering models, as it does not influence the result.

32

O. Pavlenko et al.

Fig. 4 Graph of enterprise profit dependence on combinations of input parameters values for each series of experiments

The estimated index (Fig. 4) was calculated on the obtained regression models. For this purpose, data of influenced parameters are used at maximal and minimum values of variables. It can be seen from curves that all four options have positive values only for the first, sixth, tenth, twelfth, thirteenth, and fourteenth series, that is, for the minimum value of goods delivery time. The largest profit is made by the option—“Route 4,” equal to UAH 10,211.28 for sending the corresponding type of goods, but this case is separated due to using the fastest type of transport in trunk connection— air transport. Positive profit values were compared (Table 6) to determine the most effective options from the first three routes. It is visible from calculations results (Table 6) that the option “Route 2” (seaport of Guangzhou, China—Odessa, a seaport of Ukraine—terminal of Kharkiv) highest profit values compared to the other two options: profits of “Route 1” is more by 279 UAH/kg and for “Route 3”—by 355 UAH/kg.

5 Conclusion A methodology was designed that included the following elements: development of alternatives (routes) for goods delivery; determining an evaluation parameter for selecting an effective option; building models of options for goods supply; modeling delivery process, and determining an estimate; selection of effective option according to specific criteria.

Values of influence parameters Delivery order Transportation quantity of the volume in the relevant item current batch type (kg) (kg) 1.74 260 140 260 140 260 140 1258 1.74 260 140 1258 Delivery time (h) 336 336 336 336 336 336

Table 6 Results of comparative effect determination

Compare routes 1 and 2 134.36 -279.04 -279.04 -59.48 134.36 -59.48

Effect, UAH

Compare routes 1 and 3 91.57 76.36 76.36 16.48 91.57 16.48

Compare routes 2 and 3 -42.79 355.40 355.40 75.96 -42.79 75.96

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route. . . 33

34

O. Pavlenko et al.

The routes are formed considering the existing Silk Way—on the elements of which the possibilities of transportation by rail are taken into account: ChongqingWarsaw route; by sea: Port of Guangzhou—Port of Odessa (Port of Koper); by air service Guangzhou Airport—Boryspil Airport. Simulation modeling was chosen as a research method at this stage. This type of simulation allows the complete description of relationships between the research object’s elements, considering the intensity of operations and order receipt, and probabilities of failures and risks impact. A full-factor simulation plan for four exposure parameters was developed, consisting of 16 series of experiments. As an experiment result, enterprises’ profit was obtained using initial and final values of the unit of the good according to proposed options. The regression model was found in linear form based on a regression analysis of experiment results. Each coefficient in this model indicates the influence levels of corresponding factors on the resulting parameter. This model is the most adequate because the R-square value for the first three options is near 1, and the fourth is 0.795. The effect determination results showed that “Route 2” received the highest profit values (Guangzhou seaport, China—Odessa seaport, Ukraine—Kharkiv terminal). When there was compared to “Route 1” (Guangzhou railway station, China— Chongqing railway station, China—Warsaw railway station, Poland—Kharkiv terminal, Ukraine) value is more by 279 UAH/kg, and when comparing with “Route 3” (Guangzhou seaport, China—Koper, Slovenia—terminal Kharkiv, Ukraine)—by 355 UAH/kg. These values are achieved with the maximum level of the delivery order quantity of relevant goods type, the minimum value of transportation quantity of goods in the current batch, and the delivery time. Acknowledgments This paper has been written with the support of the H2020 project “A Policy Tool Kit for the Promotion of Intercultural Competence and Diversity Beliefs, Reduction of Discrimination and Integration of Migrants into the Labor Market”, acronym FairFuture, Nr. 870307.

References 1. Kirichenko AI (2012) The application of information technologies in the management of cargo delivery processes. Transport problems: collection of scientific papers. vol. 9, NTU, pp 17–27 2. Zelikov VA, Akopova ES, Pilivanova EK, Popova LK (2019) Model of management of the risk component of intermodal transport: information and communication technologies of transport logistics. Perspectives on the use of new information and communication technology (ICT) in the modern economy. ISC 2017. Advances in intelligent systems and computing, vol 726. Springer, Cham 3. Order of the Cabinet of Ministers of Ukraine on transport strategy 2030. https:// www.kmu.gov.ua 4. Geographical structure of foreign trade in goods in January–July 2019. http:// www.ukrstat.gov.ua/operativ/operativ2019/zd/ztt/ztt_u/ztt0719_u.htm

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route. . .

35

5. Omelianenko S, Kondratenko Y, Kondratenko G, Sidenko I (2019) Advanced system of planning and optimization of cargo delivery and its IoT application. In: 3rd international conference on advanced information and communications technologies (AICT). AICT, Lviv, pp 302–307. https://doi.org/10.1109/AIACT.2019.8847744 6. Perakovi´c D, Periša M, Sente RE (2019) Information and communication technologies within industry 4.0 concept. In: Ivanov V et al (eds) Advances in design, simulation and manufacturing. DSMIE-2018, Lecture notes in mechanical engineering. Springer, Cham, pp 127–134. https://doi.org/10.1007/978-3-319-93587-4_14 7. Karabegovi´c I, Turmanidze R, Daši´c P (2020) Robotics and automation as a foundation of the fourth industrial revolution - industry 4.0. In: Tonkonogyi V et al (eds) Advanced manufacturing processes. InterPartner-2019, Lecture notes in mechanical engineering. Springer, Cham. https://doi.org/10.1007/978-3-030-40724-7_13 8. Fernando E, Surjandy, Warnars HLHS, Meyliana, Kosala R, Abdrachman E (2018) Critical success factor of information technology implementation in supply chain management. In: Literature review 5th international conference on information technology, computer, and electrical engineering (ICITACEE). IEEE, Piscataway. https://doi.org/10.1109/ ICITACEE.2018.8576979 9. Shramenko N, Muzylyov D, Shramenko V (2020) Model for choosing rational technology of containers transshipment in multimodal cargo delivery systems. In: Karabegovi´c I (ed) New technologies, development and application III. NT 2020, Lecture notes in networks and systems. Springer, Cham, pp 621–629. https://doi.org/10.1007/978-3-030-46817-0_72 10. Volkov V, Taran I, Volkova T, Pavlenko O, Berezhnaja N (2020) Determining the efficient management system for a specialized transport enterprise. Naukovyi Visnyk Natsionalnoho Hirnychoho Universytetu 4:185–191 11. Konovalenko I, Ludwig A (2019) Event processing in supply chain management – the status quo and research outlook. Comput Ind 105:229–249. https://doi.org/10.1016/ j.compind.2018.12.009 12. Borgi T, Zoghlami N, Abed M, Saber Naceur M (2017) Big data for operational efficiency of transport and logistics: a review. In: 6th IEEE international conference on advanced logistics and transport (ICALT). https://doi.org/10.1109/ICAdLT.2017.8547029 13. Delivery of cargo from China. https://sigma-logistics.com.ua/services 14. Delivery of goods by rail from China. https://fialan.ua/services/gd-dostavka-iz-kitaya/ 15. Delivery of cargo by air. https://www.china-cargo.in.ua 16. Seo Y, Chen F, Roh SY (2017) Multimodal transportation: the case of laptop from Chongqing in China to Rotterdam in Europe. Asian J Ship Log 33(3):155–165. https://doi.org/10.1016/ j.ajsl.2017.09.005 17. Kundu T, Sheu J-B (2019) Analyzing the effect of government subsidy on shippers’ mode switching behavior in the belt and road strategic context. Transp Res E 129:175–202. https:// doi.org/10.1016/j.tre.2019.08.007 18. Liu X, Zhang K, Chen B, Zhou J, Miao L (2018) Analysis of logistics service supply chain for the one belt and one road initiative of China. Transp Res E 117:23–39. https://doi.org/10.1016/ j.tre.2018.01.019 19. Jin CF, Yang HM, Ling L (2010) Wang research on optimization and debugging simulation model of logistics center based on neural network. Appl Mech Mater 38:1060–1063. https:// doi.org/10.4028/www.scientific.net/AMM.37-38.1060 20. He W, Lu T, Yu CQ (2014) A novel traffic flow forecasting method based on the artificial neural networks and intelligent transportation systems data mining. Adv Mater Res 842:708– 711. https://doi.org/10.4028/www.scientific.net/AMR.842.708 21. Subbotin SO, Oliynyk AO (2014) Neural networks: teach. Manual. ZNTU, Zaporizhzhya 22. Muzylyov D, Shramenko N (2020) Blockchain technology in transportation as a part of the efficiency in industry 4.0 strategy. In: Tonkonogyi V et al (eds) Advanced manufacturing processes. InterPartner-2019, Lecture notes in mechanical engineering. Springer, Cham, pp 216–225. https://doi.org/10.1007/978-3-030-40724-7_22

36

O. Pavlenko et al.

23. Wang L, Zhu XN, Xie ZY (2011) Object-oriented petri net modeling and analysis of China railway container freight yard logistic system. Key Eng Mater 467:990–995. https://doi.org/ 10.4028/www.scientific.net/KEM.467-469.990 24. Shramenko N, Pavlenko O, Muzylyov D (2020) Logistics optimization of agricultural products supply to the European union based on modeling by petri nets. In: Karabegovi´c I (ed) New technologies, development and application III, Lecture notes in networks and systems. Springer, Cham, pp 596–604. https://doi.org/10.1007/978-3-030-46817-0_69 25. Zhong WZ, Fu XQ, Wang YP (2013) Petri net modeling: container terminal production operation processing system analysis. Appl Mech Mater 409:1320–1324. https://doi.org/ 10.4028/www.scientific.net/AMM.409-410.1320 26. Pavlenko O, Velykodnyi D, Lavrentieva O, Filatov S (2020) The procedures of logistic transport systems simulation into the petri nets environment. CEUR Workshop Proc 2732:854– 868 27. Rossolov A, Kopytkov D, Kush Y, Zadorozhna V (2017) Research of effectiveness of unimodal and multimodal transportation involving land modes of transport. Eastern-Eur J Enterpr Technol 5(89):60–69. https://doi.org/10.15587/1729-4061.2017.112356 28. Turpak SM, Taran IO, Fomin OV, Tretiak OO (2018) Logistic technology to deliver raw material for metallurgical production. Naukovyi Visnyk Natsionalnoho Hirnychoho Universytetu 1:162–169. https://doi.org/10.29202/nvngu/2018-1/3 29. Litvinova Y, Nosal-Hoy K, Solecka K, Taran I (2020) Improvement of efficiency of processes of mining product processing at transport hubs. Naukovyi Visnyk Natsionalnoho Hirnychoho Universytetu 1:141–145. https://doi.org/10.33271/nvngu/2020-1/141 30. Luscinski S, Ivanov VA (2020) Simulation study of industry 4.0 factories based on the ontology on flexibility with using FlexSim software. Manage Prod Eng Rev 11(3):74–83. https://doi.org/ 10.24425/mper.2020.134934 31. Medvediev I, Muzylyov D, Shramenko N, Nosko P, Eliseyev P, Ivanov V (2020) Design logical linguistic models to calculate necessity in trucks during agricultural cargoes logistics using fuzzy logic. Acta Logist 7(3):155–166. https://doi.org/10.22306/al.v7i3.165 32. Steelant J (2012) Pioneering in hypersonic transportation: long term perspectives and technological challenges. In: Kontis K (ed) 28th international symposium on shock waves. Springer, Berlin, pp 39–43. https://doi.org/10.1007/978-3-642-25688-2_6 33. Silva JB, Giannotti MA, Larocca APC et al (2017) Towards a spatial data infrastructure for technological disasters: an approach for the road transportation of hazardous materials. GeoJournal 82:293–310. https://doi.org/10.1007/s10708-015-9680-0 34. Mieczkowski B (1980) Technological change in transportation in Eastern Europe. In: Mieczkowski B (ed) East European transport regions and modes. Developments in transport studies. Springer, Dordrecht, pp 282–316. https://doi.org/10.1007/978-94-009-8899-6_13 35. Medvediev I, Sakno O, Moisia D, Kolesnikova T, Rogovyi A (2020) Linear and non-linear wheel slip hypothesis in studying stationary modes of a double road train. In: Proceedings 2020 IEEE 15th international conference on computer sciences and information technologies (CSIT), vol 1. IEEE, Piscataway, pp 183–187 36. Kliuiev S, Medvediev I, Soroka S, Dubuk V (2020) Development of the intelligent rail vehicle control system. In: Proceedings 2020 IEEE 15th international conference on computer sciences and information technologies (CSIT), vol 1. IEEE, Piscataway, pp 369–372 37. Vojtov V, Kutiya O, Berezhnaja N, Karnaukh N, Belyaeva O (2019) Modeling of reliability of logistic systems of urban freight transportation taking into account stream loading. Eastern-Eur J Enterpr Technol 7(4):15–21. https://doi.org/10.15587/1729-4061.2019.175064 38. Grabis J, Haidabrus B, Protsenko S, Protsenko I, Rovna A (2019) Data science approach for it project management. Vide Tehnol Res 2:51–55. https://doi.org/10.17770/etr2019vol2.4163 39. Dobrotvorskiy S, Basova Y, Dobrovolska L, Sokol Y, Kazantsev N (2020) Big challenges of small manufacturing enterprises in industry 4.0. In: Ivanov V, Trojanowska J, Pavlenko I, Zajac J, Perakovi´c D (eds) Advances in design, simulation and manufacturing III, vol 1. IEEE, Piscataway, pp 118–127. https://doi.org/10.1007/978-3-030-50794-7_12

Mathematical Modeling as a Tool for Selecting a Rational Logistical Route. . .

37

40. Li F, Zhu YP, Wu HR (2013) Modeling and optimization of traceability system for agriculture products supply chain. Adv Mater Res 605:574–579. https://doi.org/10.4028/ www.scientific.net/AMR.605-607.574 41. Qu JH, Yao XS, Ying JL (2012) Agricultural products logistics operational pattern based on information center. Adv Mater Res 363:1679–1683. https://doi.org/10.4028/ www.scientific.net/AMR.361-363.1679 42. Vendrell-Herrero F, Bustinza OF, Parry G, Georgantzis N (2017) Servitization, digitization and supply chain interdependency. Ind Mark Manag 60:69–81. https://doi.org/10.1016/ j.indmarman.2016.06.013 43. Xue L, Zhang C, Ling H, Zhao X (2013) Risk mitigation in supply chain digitization: system modularity and information technology governance. J Manag Inf Syst 30:325–352. https:// doi.org/10.2753/MIS0742-1222300110 44. Shramenko N, Muzylyov D, Shramenko V (2020) Methodology of costs assessment for customer transportation service of small perishable cargoes. Int J Bus Perform Manag 21(2):132–148. https://doi.org/10.1504/IJBPM.2020.10027632 45. Muzylyov D, Shramenko N (2020) Mathematical model of reverse loading advisability for trucks considering idle times. In: Karabegovi´c I (ed) New technologies, development and application III. NT 2020. Lecture notes in networks and systems, vol 128. Springer, Cham, pp 612–620. https://doi.org/10.1007/978-3-030-46817-0_71 46. Muzylyov D, Shramenko N, Shramenko V (2020) Integrated business-criterion to choose a rational supply chain for perishable agricultural goods at automobile transportations. Int J Bus Perform Manag 21(2):166–183. https://doi.org/10.1504/IJBPM.2020.10027634 47. Chinese imports: how Ukrainian businesses protect their interests. https://biz.liga.net/ ekonomika/all/opinion/kitayskiy-import-kak-ukrainskomu-biznesu-zaschitit-svoi-interesy 48. China has become the largest business partner of Ukraine - infographic. https://nv.ua/ biz/economics/torgovlya-s-kitaem-kitay-stal-glavnym-delovym-partnerom-ukrainy-novostiukrainy-50048806.html 49. Ukraine-China. Colonial imbalance. https://tyzhden.ua/Economics/233713 50. Building the silk road. http://investasianmain.gelderbauerltd.netdna-cdn.com/wp-content/ uploads/2015/02/MapChinaNewSilkRoad.jpg

Commercializing M2M eSIM Networks Bassem Ali Abdou

Abstract This paper illustrates the different approaches to commercialize M2M eSIM networks. It will first highlight cellular telecommunication position in the IOT connectivity landscape and SIM evolution to the digital eSIM form, with basic information about architecture, interfaces, and operations; the paper illustrates the three main variances of M2M eSIM networks based on the ownership of the secure route element of the subscription manager: Operator controlled variance (in which the OEM outsources the secure route to a network operator), OEM controlled variance (in which the OEM acquires and operates the secure route), and M2M service provider controlled variance (in which the OEM outsources the secure route to a M2M service provider). Mobile telecommunication regulations and business conditions led to emergence of two approaches: lent eSIM and eSIM ownership transfer. Ecosystem players should consider available resources, limitations, advantages, and disadvantages of each variance before selecting the commercialization approach. The complex mesh network and long list of standardized and nonstandardized interfaces among ecosystem building blocks present a challenge to M2M eSIM technology adoption. The idea of having independent hub entities to interconnect subscription manager platforms and connectivity managers from the different players in the globe presents future research, standardization, and business opportunity that can address M2M eSIM network complexity and time-to-market. Keywords eSIM · eUICC · EUM · GSMA · M2M · Subscription manager

B. A. Abdou () Mobily, Riyadh, Saudi Arabia © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 D. Cagáˇnová, N. Horˇnáková (eds.), Industry 4.0 Challenges in Smart Cities, EAI/Springer Innovations in Communication and Computing, https://doi.org/10.1007/978-3-030-92968-8_3

39

40

B. A. Abdou

1 Introduction The IOT is about connecting devices having embedded software and sensors to the internet in a way that enables information gathering, analytics, decision making, and taking actions; therefore, it presents endless opportunities in almost all fields [1]. Connectivity is quite important part of IOT as it has its implications on IOT service, business process, and cost [2]. Figure 1 illustrates the different network technologies that can be used for IOT networks. To the left, we have WPANs (Wireless Personal Area Networks), which use technologies like Zigbee (standardized under IEEE 802.15.4) and WiFi (standardized under IEEE 804.11). Moving to the right, we have the WANs (wide area networks), where licensed mobile telecommunication technologies such as cellular 2G, 3G, 4G, 5G, CAT-M (category-M), and NB-IOT (narrow band IOT) exist; also unlicensed LPWAN (low power wide area network) technologies such as LoRa and SigFox exist. Many researches addressed those technologies, associated technical differences and business implications [2–4]. Among the different alternatives, cellular technologies are quite advantageous when availability, coverage, reliability, and speed are needed. In fact, telematics industry, which is mainly concerned with vehicle communication, navigation, tracking, positioning, and status monitoring, is highly dependent on cellular telecommunication in specific [5]. Currently, NB-IOT and CAT-M see massive IOT adoptions and compete against unlicensed rivals; also, 5G technology is expected to address wider stream of connections and applications and may connect more than four billion cellular devices by 2024 [6]. All cellular technologies necessitate the existence of a SIM (subscriber identity module) in the communication module part

Fig. 1 IOT network technologies

Commercializing M2M eSIM Networks

41

of the IOT device, in order to host a network operator profile and utilize some mobile network operator in transferring data from device to the internet and vice versa. The use of traditional removable SIMs in M2M business has two operational challenges: (1) environmental conditions for M2M are not as controlled as they are in case of mobile phones, with which traditional removable SIMs can be damaged, (2) reaching M2M devices for swapping SIMs when needed (e.g. to change the mobile network operator) can be very difficult. GSMA (GSM Association) responded to those challenges by putting new standards for a new digital kind of SIM called eSIM (Embedded SIM). In eSIM technology, physical SIM chipsets are built-in inside devices while operator profiles are hosted on the cloud; the ecosystem provides industry players with remote provisioning capabilities to push, activate, deactivate, and delete operator profiles; this new technology eliminates the use of the less-secure idea of soft SIMs [7]. eSIM is a transformational step in digitizing the mobile telecommunication world and making it more appealing to the new IOT era for which cloudification and seamless automation tend to be prerequisites. Among the different IOT verticals, eSIM is particularly important for the automotive sector; eSIM does not only address technical challenges like mobility and coverage [8] but also helps car manufacturers to fulfill eCall service mandate [9] without the expensive—and sometimes forbidden—permanent roaming in foreign networks. GSMA’s remote provisioning architecture for eUICC (embedded universal integrated circuit card) technical specification version 4.2 [10] is the latest one for M2M eSIM (issued in July 2020); however, many of the live deployments at the time of writing this paper are still following GSMA’s older release, which is 3.2 [11]. GSMA’s M2M eSIM track facilitates control capabilities for the enterprise to remotely manage devices’ connectivity in the field and is different from the consumer eSIM track; the latter has its own GSMA standard [12], which is relevant when control needs to be at the consumer side; consumer track is beyond the scope of this document. eSIM has been addressed in the literature from different angles, including the adoption for enabled devices in specific markets [13], generic impact on consumer market and telcos [14], business process [15], and categories of use cases [16]. Practice and ecosystem for consumer eSIM have been addressed in detail [17]. Practice and ecosystem for M2M have been also addressed in a valuable paper back in 2015 [18] but the assumption that eSIM data preparation part was not controlled by mobile network operator contradicted with actual implementations, how business ecosystem evolved in real life and how it can be managed; what is more, evolution of standards created clearer picture that might have not been available at that time. This paper will analyze the complex M2M eSIM networks in the light of GSMA’s remote SIM provisioning standard releases 3.2 and 4.2, and investigate how the building blocks can be distributed among ecosystem players, the part that is not addressed explicitly in the standards and in the literature. While basic network information will be illustrated based on the standard specifications, the intention is not to provide detailed and complete descriptions for components, interface protocols, functions, and parameters, which are already available in GSMA specification releases. The Intention, however, is to clarify the main new

42

B. A. Abdou

eSIM network building blocks, interfaces, and operations that did not exist in physical SIMs world; this will be mandatory to derive the different approaches to commercialize M2M eSIM networks and the different ways players can join the ecosystem network in each approach. Evidences from the market will be provided whenever available and the limitations of the existing eSIM network topology will be also addressed.

2 M2M Remote SIM Provisioning Overview 2.1 Basic Network Architecture Figure 2 illustrates GSMA Architecture for the M2M eSIM remote SIM provisioning system (RSP). The architecture introduces three standard elements that are new to the mobile telecommunication world and are created specifically for the M2M eSIM: • eUICC stands for embedded universal integrated circuit card. This chipset is similar to the traditional SIM (subscriber identity module) in the fact that it hosts the network operator profile; however, it is usually soldered in the M2M equipment and is structured differently in order to host multiple remotelycontrolled operator profiles. • SM-DP stands for subscription manager—data preparation part; it is the element responsible for storing the operator’s eSIM profiles and for pushing those profiles towards M2M devices that are equipped with the eUICC. SM-DP is assumed to be owned and controlled by operator. This is irrelevant to the fact that—like any network element—SM-DP can be purchased from a telecom vendor or even acquired as a service on the cloud. • SM-SR stands for subscription manager—secure route part; it is the element responsible for the secure communication with the eUICC; basically, it stores EIS (eUICC Information Set) and is the gatekeeper that allows eSIM profile delivery to the eUICC. Operator is basically the MNO (mobile network operator) or MVNO (mobile virtual network operator) owning the profile and network; it is usually presented in the actual ecosystem by the CM (connectivity manager). M2M SP stands for machine-to-machine service provider, to which MNO can delegate eSIM operation management. EUM stands for eUICC manufacturer, which is the entity that fabricated the eSIM chipset before it was installed in the M2M device (a car for instance). CI stands for the certificate issuer, which is needed to secure the ecosystem by issuing relevant certificates.

Commercializing M2M eSIM Networks

43

Fig. 2 M2M eSIM remote provisioning system [10]

2.2 eUICC Architecture Separation between the physical chipset and the operator profile was the core part of the evolution from physical SIM to eSIM. The communication module of any M2M device gets produced with built-in chipset that has an initial bootstrap profile for initial communication with the M2M application; in a later stage, another new profile could be pushed to the chipset over the air and could be activated instead of the bootstrap one. The chipset can then have more than one profile. As shown in Fig. 3, three elements exist inside the eUICC: • ISD-R stands for issuer security domain-root and is corresponding to an SM-SR. • ECASD stands for eUICC certificate authority security domain. It is installed on the chipset by the EUM while manufacturing. • ISD-P stands for issuer security domain—profile, and is corresponding to an SMDP. Multiple ISD-Ps can exist on an eSIM, with each one hosting an operator profile. Only one can be active at any given time.

44

B. A. Abdou

Fig. 3 eUICC architecture [10]

2.3 Interfaces Excluding ES6 interface, the architecture illustrated in Fig. 2 includes new interfaces that have been introduced by GSMA particularly for remote M2M eSIM: • ES1 is the interface between EUM and SM-SR; it is used for uploading eUICC information on the SM-SR. • ES2 is the interface between the MNO or MVNO owning the profile and the SM-DP platform hosting the eSIM profile information. It is used for profile management functions such as pushing a profile, disabling or deleting it. • ES3 is the interface between the SM-DP and SM-SR; it is used in ISD-P creation and deletion and also in profile operations such as enable and disable; notifications go through this interface in the opposite direction (from SM-SR to SM-DP). • ES4 is the interface between the MNO and SM-SR; excluding profile push, it can be used for the different profile management operations similar to what is happening through ES2-ES3 interfaces. • ES4a has been introduced in the recent GSMA release 4 and is used by the operator to allow eSIM PLMA (profile life cycle management authorization) for a M2M player. • ES5 is the interface between SM-SR and eUICC; it can be based on SMS, HTTPs or CAT-TP (card application toolkit transport protocol); it is used by the SM-SR to create an ISD-P on the eUICC, enable, disable a profile on an ISD-P or even delete the ISD-P; it is also used for the relevant notifications in the opposite direction from the eUICC to the SM-SR. • ES6 is the interface between the MNO (or MVNO) and its profile on the eUICC; this is the traditional OTA (over the air) interface that has been used by MNOs for several years to do regular changes on the SIM such as SMSC (SMS Center)

Commercializing M2M eSIM Networks

45

address change. This interface is not new and is standardized by ETSI, not GSMA [19, 20]. • ES7 is the interface between two different SM-SRs and is used only in case of changing the eUICC ownership from a business entity (presented by specific SM-SR) to another one (presented by another SM-SR). Ownership change will be further discussed in Sect. 5. • ES8 is the interface between SM-DP and its dedicated domain on the eUICC; it is used for downloading an eSIM profile from SM-DP to the eUICC and installing it. ES8 communication is tunneled through ES3 and ES5.

3 M2M eSIM Main Operations Bootstrap profile that is initially loaded on the chipset is mandatory in order for the device to start communicating with the ecosystem [21]. In a later stage of the device lifecycle, the OEM, network operator or M2M entity (depending on the setup variance as will be illustrated in Sect. 4) may decide to change the working profile in the device; this can happen when, for example, a car manufactured in a certain country with a bootstrap profile (typically from some local operator) gets exported to another country where another operator’s profile (destination operator) should be pushed and activated. This section will shed some light on two main network operations that were introduced to the mobile telecommunication world as part of M2M eSIM technology: eUICC ordering (with bootstrap profile loading included) and new profile loading.

3.1 eUICC Ordering Figure 4 illustrates a simplified version of the eUICC ordering process. An agreement is assumed to be in place between OEM and some MNO or MVNO (an entity that has an operator profile), specifying commercial and technical terms. The OEM orders eUICCs from the operator; the operator then relays the request to the EUM, specifying the OEM to which eUICC chipsets will be shipped and the SMSR to which EIS data should be passed. The EUM runs the personalization process in order to prepare the chipsets, with the right operator profile pre-loaded and then takes three main actions: • Deliver subscription data back to the operator. • Ship the physical eUICCs to the OEM. • Update the SM-SR with EIS data. The operator will have to provision the eSIM profiles in the different network elements in order for profiles to be known to the mobile network. The OEM will have to embed the physical eUICC in the device. Also, EUM will have to use ES1

46

B. A. Abdou

Fig. 4 Simplified eUICC ordering process

interface for storing the EIS information in the right SM-SR; this SM-SR will be the gateway towards the eUICC and the business entity owning it will be the one assuming control on the chipset and its profile operations. It should be highlighted that for simplicity, OEM is presented by one box in the illustrated process; in practice, there might be a business entity for manufacturing the DCM (data communication module) with eSIM embedded and another business entity (for example, a car manufacturer) integrating the DCM in the final product.

3.2 New Profile Loading and Activation With some low-level details eliminated, Fig. 5 summarizes M2M eSIM profile download and activation procedure in an abstract way (i.e. regardless of the business entity owning or operating the technical function); new profile loading can be triggered when, for example, a car lands in a new country and there is a need to push a local profile from an operator covering that country. The procedure has three main stages: • ISD-P Creation. • Profile Download.

Commercializing M2M eSIM Networks

47

Fig. 5 Simplified M2M profile download process

• Profile Enablement. ISD-P creation stage starts when the operator owning the eSIM profile signals a download profile command to its SM-DP through ES2 interface; based on the EID (eUICC ID) and SM-SR ID received, SM-DP will retrieve the eUICC information set from the right SM-SR and will then instruct it to create an ISD-P in the eUICC. Following the mutual authentication, profile download stage starts when SM-DP initiates send data command towards SM-DR, with the profile information secured and tunneled to the right ISD-P through the SM-SR and ISD-R; the process can be triggered multiple times until profile data is complete then operator gets updated through ES2 interface. In the profile enablement stage, command is passed through ES3 interface to SM-SR, which in turn contacts the ISD-R through ES5 interface for the sake of the corresponding ISD-P enablement. eUICC does a quick restart, re-attach to the network using the right profile and acknowledgment is sent back to the SM-DP and operator accordingly.

48

B. A. Abdou

4 Ecosystem Variances A particular eSIM network variance refers to a certain distribution of the ecosystem building blocks across ecosystem players: network operators, OEM and M2M service provider (M2M-SP), and with the interfaces between those players. The following rules present the framework based on which the research is made: • SM-DP is logically positioned within the boundaries of the telecom network operator (MNO or MVNO). • Operator domain can be an MNO or MVNO; most importantly, this entity has to have its own eSIM profile. If a business entity has its own profile and proposes itself in the market as a machine-to-machine player, then it will be still considered as an operator in this paper’s classification. An M2M-SP is not supposed to have its own SIM profile. • SM-SR can be logically positioned within the boundaries of any player in the ecosystem: operator, OEM or M2M-SP. Having it as a cloud service hosted by some vendor does not change the fact that it is controlled by the operator, OEM or M2M player in terms of business rules and integrations. Classification will be based mainly on the entity owning the SM-SR, the controlling gateway towards the eUICC. • Having SM-DP or SM-SR as a cloud service hosted or operated by some vendor does not change the fact that in terms of business and technical integrations, it is controlled by the acquiring network entity (operator, OEM or M2M). Therefore, by considering all possible combinations, it is hypothesized that M2M eSIM network commercialization can be on the form of any of the three variances as shown in Table 1. The next parts will illustrate the three variances in the light of the hypothesis and previous assumptions, knowing that each business entity is assumed to have a CM in order to deploy the business logic and mediate the non-standard technical relationship with other entities; this is a common practice among ICT players. Sharing prerequisite data for initiating M2M GSMA-standard operations (such as EID) is assumed to happen between the CMs through inter CM interface (named as CMI in this paper), after which data is passed to GSMA-Standard elements (SM-DP, SM-SR, and eUICC). Table 1 M2M network variances classification framework SM-DP owner Network operator Network operator Network operator

SM-SR owner Network operator OEM M2M-SP

Variance Operator controlled SM-SR OEM controlled SM-SR M2M-SP controlled SM-SR

Commercializing M2M eSIM Networks

49

4.1 Operator Controlled SM-SR Variance In an operator controlled SM-SR variance, the OEM relies on an MNO that acquires and operates complete subscription manager (SM-SR and SM-DP) and typically provides a bootstrap profile. As shown in Fig. 6, the MNO’s connectivity manager (CM-OP-1) typically sits on top of the SM-DP and SM-SR; it can decide if another profile is needed (for example, when the vehicle moves to a new country) and can trigger the request to the right operator’s connectivity manager (CM-OP-2) through the non-standard interface CMI-2 (assuming agreement and proper integration are in place); operator-2 can in turn initiate profile download through its SM-DP (SMDP-2) and its ES3 interface with the SM-SR. This variance follows a traditional business function outsourcing model, where an OEMs hires some telecom operator to implement and manage this kind of network business. Operator controlled variance is easier for OEMs and car manufacturers that do not prefer deep involvement in setting and managing the telecommunication part of the business and would like to avoid upfront investments but it implies

Fig. 6 Operator controlled SM-SR variance

50

B. A. Abdou

heavy dependency on a single operator. This variance can be also advantageous to operators that will not only have new revenue stream from OEMs, but also will benefit from the full control of the subscription manager elements (SM-DP and SMSR) and eliminate some interoperability risks. Operators can synergize if the same service can be provided to multiple OEMs; also, group operators with presence in several countries will find this approach attractive. On the negative side, the operator is likely to face a scalability issue while establishing and managing the links with other operators as number of destination countries grows; OEMs will want to secure local connectivity in the different countries and MNO will need to establish and manage bilateral interconnects with many SM-DP platforms from many operators over the globe, which can extend the TTM (Time-To-Market) significantly and can be heavy operational load on system and staff. Some press releases in industrial countries such as Japan [22] and China [23] suggest that complete subscription managers (i.e. SM-DP and SM-SR) have been acquired by the MNOs and that served (or to-be-served) OEMs would subsequently need no SM-SR. Similarly, a press release from the MVNO cubic telecom [24] stated that Skoda cars would use cubic telecom’s SM-SR for eSIM service enablement. What is more, according to GSMA SAS Accredited Sites [25], each of the telecom operators Reliance (India) and Telenor (Norway) has both SM-SR and SM-DP, which indicates that they have this kind of setup or at least considered it at a certain stage.

4.2 OEM Controlled SM-SR Variance In an OEM controlled SM-SR variance, the OEM is stepping in the telecommunication business by acquiring the SM-SR part of the subscription manager and integrating it with external MNOs or MVNOs. As shown in Fig. 7, the OEM typically deploys a connectivity manager (CM-OEM) and has to be present in the technical and business relationship with the different operators. Integrations with data preparation platforms from other operators looks unavoidable in this variance and it would be mandatory to add more integrations if the device moves to a different country where permanent roaming is not practical or not allowed. The fact that until the time this paper was prepared in 2020, GSMA certified sites has no automotive OEMs [25] indicates that OEMs that decided to acquire an SM-SR followed an Application Service Provider (ASP) kind of outsourcing. Profile operations are initiated between connectivity managers of the different entities using the non-standard interfaces (CMI-n) and then passed to standard elements (SM-DPs and SM-SRs) using the standard interfaces. This approach eliminates the dependency of vehicle manufacturer on a main telecommunication player and maintains eUICC control in the hands of OEM. It will, however, necessitate some upfront investment and existence of telecommunication team in the OEM workforce in order to manage SM-SR work, associated

Commercializing M2M eSIM Networks

51

Fig. 7 OEM controlled SM-SR variance

CM-OEM integration, and interfaces with other operators’ SM-DP platforms; such kind of mobile network integration and operation work can present new skillset that OEMs in favor of this model will have to acquire, at a cost that may be understood only if the right economy of scale exists. Similar to the operator in the MNO controlled variance, the OEM is likely to face the scalability issue associated to the mesh kind of networks in this approach when numbers of destination countries and peer operators grow. On the other side, and although M2M eSIM is still new revenue stream for operators, they will perceive the OEM controlled variance as an attempt to commoditize their connectivity offering. A press release from the MVNO Transatel [26] indicated that the OEM Scania and the Transatel followed this approach; Transatel’s eSIM profiles were to be transferred from Transatel’s SM-DP to Scania trucks through Scania SM-SR. Another release from Idemia [27] recommends that Fiat Chrysler followed similar approach by acquiring a subscription manager platform.

52

B. A. Abdou

4.3 M2M-SP Controlled SM-SR Variance It is technically possible for an M2M service provider to acquire or build an SMSR and connectivity manager (CM-M2M), aggregate profiles from several operators without having its own eSIM profile (shown in Fig. 8), and bridge the relationship between the OEM and mobile operators. Neither clear press release nor publicly available information indicated commercial implementation of this variance. As expected, profile operations are initiated between connectivity managers of the different entities using the non-standard interfaces (CMI-n) and then passed to standard elements (SM-DPs and SM-SRs) using the standard interfaces. Like the operator controlled variance, traditional telecommunication function outsourcing from the OEM to the M2M-SP should take place in this variance, but unlike the operator controlled variance, the hired M2M-SP should be more agile than traditional operators, free of group commitments, and neutral while picking up operator profiles. Potential challenge for this approach can be for the M2M business entity to play that role with the absence of own profile, especially when it comes to

Fig. 8 M2M controlled SM-SR variance

Commercializing M2M eSIM Networks

53

bootstrap stage, where the M2M business entity will have to setup a deal with an operator to use its profile for this purpose. Scalability will be still a major challenge for the M2M-SP as in addition to profile management, establishing and managing the large number of bilateral interconnects are unavoidable. with lack of publicly known reference for this model, it remains unproven; however, it may present a business opportunity for startups that do not own profiles but are technically capable of managing the complex integrations and operations with different MNOs and willing to sell neutrality as main advantage to OEMs.

5 Ecosystem Challenges and Solutions Like any new technology, M2M eSIM technology comes with its challenges and limitations that are to be resolved as technology matures. This section will cover three points that present business and regulatory challenges to the industry and available or recommended solutions that might be adopted by ecosystem players.

5.1 MNOs with no SM-DP Some destination operators would prefer to fast-track their go-to-market plans to enter the M2M eSIM market and get their profiles pushed to imported vehicles before having their SM-DP ready; they can basically do so by lending some profiles to some other operator that owns a SM-DP platform, runs M2M eSIM business, and is ready to consider those foreign eSIM profiles while deciding which profile is to be pushed to the device. It is usually temporary solution that gets adopted until the destination operator builds its own SM-DP. As shown in Fig. 9, if profiles are to be lent to an MNO or MVNO, then setup would be a special case of the operator controlled variance, with one main difference: SM-DP-1 will host profiles not only from operator-1, but also from the lending operators. Its connectivity manager platform (CM-OP-1) would then have to host and apply the logic of selecting, pushing, and activating the right profile. The lending operator would maintain partial visibility on the its profiles’ usage using its CM. Lack of public references of this variance makes it unproven; however, if implemented, it would have the advantages of fast, easy, and inexpensive setup for some destination operators; on the negative side to destination operators, decision to push the profile is completely taken by the hosting operator; also, the idea of having the profile prepared and hosted on a foreign operator’s environment is not common and is questionable from security point of view. The operator owning and operating the host SM-DP on the other side shall include the necessary business and technical operations associated to foreign profiles.

54

B. A. Abdou

Fig. 9 Lent eSIM model

5.2 eSIM Ownership Change Telecom regulations can enforce changing eSIM ownership, which would be technically interpreted as a change of the SM-SR platform that controls the eSIM. For example, some countries like Turkey enforce that ecosystem elements (including SM-SR) exist within the boundaries of the country [28]; this will necessitate the transfer of eUICC ownership from a foreign SM-SR to another local one, should any imported eSIM equipped vehicle is to stay connected. Business aspects can also lead to the same when, for instance, an OEM changes the operator or M2M player it outsourced SM-SR to, or when an OEM that owns an SM-SR considers changing its business model by outsourcing the connectivity business to MNO, MVNO, or M2M service provider. GSMA standard addressed this point by adding the ES7 interface, flow, and relevant ecosystem functions [29] as shown in Fig. 10. The initiator can be the connectivity manager of any entity, which first makes sure that the new SM-SR (SM-SR2) is willing to accept the transfer, then issues the transfer request to the working SM-SR (SM-SR1). EIS data exchange and key

Commercializing M2M eSIM Networks

55

Fig. 10 SM-SR change [29]

establishment takes place among SM-SR1, SM-SR2, and eUICC until process is completed, after which initiator and operators are updated. eSIM ownership change is a process of changing the controlling entity and is not a new variance. Simply it involves switching between any two of the three main variances (operator controlled, OEM controlled, or M2M-SP controlled), including switching between two different networks of the same variance (e.g. from operator-x controlled to operator-y controlled). Aside from the technical aspects addressed by GSMA, eSIM ecosystem players should be aware of the possibility of outsourcing changes during the service life cycle and make sure that contractual agreements and workforce planning are flexible enough to absorb such situations.

5.3 Complex Integration and Scalability Like any ecosystem with multiple players and different interfaces, interoperability is important for the M2M eSIM business to grow [30]; this will be important for both standard and non-standard interfaces. While ES4 interface was not a must-to-deploy in the light of GSMA release 3 (all its functions could be facilitated through ES2-ES3 route), ES4a looks mandatory to have in release 4 at least in order to regulate PLMA (which is an important function); the latter interface is relatively new and it is not clear yet how real life deployments would be.

56

B. A. Abdou

Fig. 11 eSIM hub high level architecture

Integration and Interactions between communication managers of the different entities (CMI interfaces in the different variances) are not standardized. Also, both GSMA standard and the global market lack eSIM hub entities that focus on scalability and on elminating the mesh network of SM-SR, SM-DP and CM platforms from different operators, OEMs and M2M players. If ES3 and CMI functions are routed through central hubs in a standard, secure, and reliable way, interconnect setup, integration, and business operations such as profile acquisition, profile activation, and deactivation would be much easier. As we may expect from any mesh to star topology network evolution, this would increase the adoption of M2M eSIM and shorten the TTM significantly for M2M eSIM business stakeholders. Simply it will be faster and cheaper for all stakeholders to enter the market and achieve global reachability and it will be easier to implement and manage one or few integration links instead of multiple ones. Figure 11 has high level architecture for this kind of setup, where a new interface (ES3 ) is needed to bridge SM-SRs and SM-DPs through the proposed eSIM hub. eSIM Hub should be also able to connect to another eSIM hub, making global connectivity even easier and faster. A typical business model to be applied by the entity owning and operating such eSIM hub would be relying on setup fees, transaction-based charges, and annual support fees. The long TTM and complex integration in the mesh setup raises the cost for operators and OEMs to adopt eSIM technology, which leaves good margin for profitability to the business entity owning and operating the eSIM hub. The hub approach is not new to the mobile telecommunication world; in fact, it contributed to the growth and scalability of SMS globally in a way that was hard to achieve if only bilateral agreements were to be relied on [31]. It was also followed

Commercializing M2M eSIM Networks

57

in other services such MMS [32] and IP Exchange (IPX) [33]. Therefore, eSIM should not be different. eSIM hubbing is a potential area where further research, standardization, and business work are needed.

6 Conclusion This paper offers a blueprint for how to commercialize M2M eSIM networks to cellular IOT industry practitioners. The entity owning the SM-SR platform assumes control on the M2M eSIM, and depending on this rule, there are three main variances that can be followed by practitioners in order to commercialize M2M eSIM services: operator controlled SM-SR variance, OEM controlled SMSR variance, and M2M-SP controlled SM-SR variance. Operator controlled and OEM controlled variances are proven as there exist public references for them; however, M2M-SP controlled variance is not proven as it lacks publicly announced evidences. M2M eSIM challenges include possible changes in eSIM ownership, which has been addressed from technical perspective by GSMA standards. The different business entities should consider their available resources, limitations, advantages, and disadvantages before deciding the selected variance. One more M2M eSIM challenge is the complex mesh network of SM-DPs, SM-SRs, and CMs. The paper addresses this challenge by recommending the enrichment of M2M eSIM standards so as to add star topology and introduce a new central function block called eSIM hub; there will be a need for a new interface between SM-DP and SM-SR nodes on one side, and the proposed eSIM hub node on the other side. Business entities need to consider building and operating this hub node, and design an associated business model.

References 1. Madakam S, Ramaswamy R, Tripathi S (2015) Internet of things (IoT): a literature review. J Comput Commun 3:164–173. https://doi.org/10.4236/jcc.2015.35021 2. Ding J, Nemati M, Ranaweera C, Choi J (2020) IoT connectivity technologies and applications: a survey. IEEE Access 8:67646–67673 3. Osman NI, Abbas I (2018) Simulation and modelling of LoRa and sigfox low power wide area network technologies. IEEE, Piscataway, pp 1–5 4. Mekki K, Bajic E, Chaxel F, Meyer F (2018) Overview of cellular LPWAN technologies for IoT deployment: Sigfox, LoRaWAN, and NB-IoT. In: 2018 IEEE international conference on pervasive computing and communications workshops (PerCom workshops). IEEE, Piscataway, pp 197–202. https://doi.org/10.1109/PERCOMW.2018.8480255 5. ASASHOP (2008) Telematics past, present and future, automotive service association report. http://www.asashop.org/news/asaresources/ASAtelematics_0508.pdf 6. Cerwall P et al. Ericsson mobility report. https://www.ericsson.com/assets/local/mobilityreport/documents/2018/ericsson-mobility-report-november-2018.pdf 7. Anonymous (2015) Understanding SIM evolution. GSMA intelligence

58

B. A. Abdou

8. Rehak A, Freire I (2019) eSIM solutions drive new opportunities for global IOT services. Ovum 9. Sealy P (2019) The true value proposition of the eSIM. ABI Research Tata Communications 10. Remote provisioning architecture for embedded UICC technical specification version 4.2. GSMA Standard. 2020 11. Remote provisioning architecture for embedded UICC technical specification version 3.2. GSMA Standard 2017 12. RSP technical specification version 2.2. GSMA Standard 2017. 13. Arifin AS, Pradipta A, Gunawan D (2017) Modelling and analysis e-SIM in Indonesia. In 15th international conference on quality in research (QiR): international symposium on electrical and computer engineering, Nusa Dua, pp 276–280. https://doi.org/10.1109/QIR.2017.8168496 14. Acker O (2016) How the disappearing SIM card will liberate the consumer and scramble telco roles. Pwc 15. Anonymous (2018) eSIM white paper, the what and how of Remote SIM Provisioning. GSMA 16. Gerpott TJ, May S (2017) Embedded subscriber identity module eSIM. Bus Inf Syst Eng 49:293–296 17. Abdou BA (2019) Commercializing eSIM for network operators. In: 2019 IEEE 5th world forum on internet of things (WF-IoT). IEEE, Piscataway, pp 616–621 18. Vesselkov A, Hammainen H, Ikalainen P (2015) Value networks of embedded SIM-based remote subscription management. In: 2015 conference of telecommunication, media and internet techno-economics (CTTE). IEEE, Piscataway, pp 1–7 19. TS 102 225- Smart Cards; Secure packet structure for UICC based applications Release 13, ETSI Standard (2018) 20. TS 102 225- Smart Cards; Remote APDU structure for UICC based applications Release 13, ETSI Standard (2016) 21. Smyth B, Quaglina E, Meyer M (2019) An overview of GSMA’s M2M remote provisioning specification 22. Gemalto Press Release. https://www.thalesgroup.com/en/markets/digital-identity-andsecurity/press-release/kddi-in-japan-selects-gemalto-s-connected-cars-and-iot-solution. Accessed 3 October 2020 23. Gemalto Press Release. https://www.thalesgroup.com/en/markets/digital-identity-andsecurity/press-release/china-mobile-enters-the-connected-car-market-with-gemalto-s-remotesubscription-management-solution. Accessed 3 October 2020 24. Cubic Telecom Press Release. https://www.cubictelecom.com/Media/PressRelease/44. Accessed 3 October 2020 25. GSMA SAS Accredited Sites. https://www.gsma.com/security/sas-accredited-sites/. Accessed 3 October 2020 26. Transatel Press Release. https://www.transatel.com/press-releases/transatel-and-gd-sign-withscania-to-provide-trucks-with-machine-to-machine-connectivity-worldwide/. Accessed 3 October 2020 27. Idemia Press Release. https://www.idemia.com/press-release/global-market-leader-fiatchrysler-automobiles-selects-idemias-connectivity-solutions-improve-connected-vehicleexperience-2020-02-10. Accessed 3 October 2020 28. Mondaq. http://www.mondaq.com/turkey/x/795466/Telecommunications+Mobile+Cable +Communications/Recent+Decision+Of+Turkish+Telecommunications+Authority+On +Remote+Programmable+SIM+Technologies. Accessed 3 October 2020 29. Embedded SIM Remote Provisioning Architecture Version 4.0, GSMA Standard (2019) 30. Makino K, Kishi D, Bian J (2018) Building of GSMA3.1-compliant eSIM Commercial System for IOT/M2M through Partnership between operators. NTT Docomo Tech J 20 31. Anonymous. Open connectivity SMS Hubbing Architecture 2.0. GSMA Association (2009) 32. Bodic GL (2005) Mobile messaging technologies and services: SMS, EMS and MMS. Wiley, London 33. Pentitinen J (2015) The telecommunications handbook, engineering guidelines for fixed, mobile and satellite systems, 2nd edn. Wiley, London

Closed Cycle of Biodegradable Wastes in Smart Cities Michal Holubˇcík, Jozef Jandaˇcka, and Juraj Trnka

Abstract The key issue of modern cities is their energy independence, which is still insufficient for most cities in the world. Outages in the supply of energy raw materials can endanger the basic functioning of cities, especially with regard to the heat supply of cities in the colder parts of our earth. This threat is exacerbated as far as the energy source is located further away from the city. It also increases transportation costs, which also increases the already high carbon footprint of the city. The solution to this problem is to find local sources of raw materials, which is the focus of this article. The most important of the energy raw materials in cities are municipal wastes. In addition, their biodegradable component, which makes up the majority, is a clean renewable resource. By incinerating this biodegradable municipal waste, we are able to reduce the cost of transporting and landfilling waste from cities and, on the other hand, obtain a source of clean and environmentally friendly alternative fuel. In this way, we create an alternative to the currently used fossil fuel coal, which often has to be imported to cities from great distances. Our article examines the benefits of using the basic components of this biodegradable waste generated in city parks, gardens or in the surrounding fields, and forests near cities. By burning these raw materials, we are able to create a closed cycle of these raw materials and ensure a balanced use of resources on our planet. Keywords Alternative biomass · Municipal waste treatment · Incineration

M. Holubˇcík () · J. Jandaˇcka · J. Trnka Faculty of Mechanical Engineering, Department of Power Engineering, University of Žilina,Žilina, Slovakia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 D. Cagáˇnová, N. Horˇnáková (eds.), Industry 4.0 Challenges in Smart Cities, EAI/Springer Innovations in Communication and Computing, https://doi.org/10.1007/978-3-030-92968-8_4

59

60

M. Holubˇcík et al.

1 Introduction to the Issue 1.1 Urban Energy Dependence All cities need energy that, like blood in a human’s veins, drives all of the city’s life functions. The energy needs of cities can be divided into three categories, such as heating, transportation, and drive of appliances. To meet the needs of cities we need to provide them with a supply of sufficient energy raw materials. Raw material sources can be transformed into individual energy carriers, which are electricity, heat, pressure, or chemical energy bound in some compounds, such as hydrogen. The conventional energy sources for our cities are mostly fossil fuels such as coal, oil, and natural gas, which we extract in great distances from cities and import to them using railways, oil, and gas pipelines. Among all the energy needs of cities, our article specializes in supplying cities with heat, and currently the most widely used source of heat in cities is still fossil coal. Not all coal is of the same quality and calorific value, and in many parts of the world they are forced to burn even smaller calorific forms of coal that are slowly approaching the calorific values of biomass [1]. In most cases, coal is mined in deep or surface mines far from human settlements. In most cases, coal is mined in deep or surface mines far from human settlements. The mined coal is transported to the place of consumption by long journeys by trains or trucks. Mining has a negative impact on the environment, but problems also arise with the transport and disposal of toxic coal ash. All life cycle of coal has a negative impact on the environment, from mining through transport to combustion. The biggest problems with burning coal in cities are dangerous emissions of sulfur in the air and toxic coal ash that is difficult to deal with. This system is not environmentally sustainable, but we also know of other solutions that can help us reverse it (Fig. 1).

Fig. 1 The current state of energy supply to the cities

Closed Cycle of Biodegradable Wastes in Smart Cities

61

1.2 Closed Cycle of Local Renewable Resources Modern urban energy supply solutions come with the use of renewable energy sources that are available locally in the urban area [2]. In addition to the use of solar, wind, hydro or geothermal energy, there is also a significant source of biomass located directly in urban green areas such as garden parks or surrounding fields and forests. Green vegetation is constantly growing and producing in addition to valued agricultural and forestry crops, also biodegradable waste, which often ends up unnecessarily rotting in a landfill or field. Local governments, together with farmers and foresters, already collect this waste regularly at present, and the new European legislation also requires its sorting, which allows us to make better use of it. The incineration of clean biomass sorted from mixed municipal waste is thus becoming the ecological fuel of the future. In addition to that, the European waste hierarchy prefers energy recovery of biodegradable waste to composting or landfilling [3]. Urban thermal power plants can thus easily exchange imported dirty coal for local waste biomass, in whole or in part [4]. The price of waste biomass is significantly cheaper than coal and short transport routes will further reduce transport costs. Season waste biomass from local forest fields and parks can be stored near local thermal power plants and gradually incinerated as needed in the colder months of the year [5]. Pure organic ash can be returned to the land from which it was taken and scattered into fields or forest areas where the cycle of this material will naturally close and give rise to new plants (see Fig. 2).

Fig. 2 Closed cycle of bio-waste treatment around cities

62

M. Holubˇcík et al.

2 Biodegradable Waste Treatment 2.1 Biodegradable Waste Statistics With the growing consumption of human society, waste is becoming a really important source of raw materials for recycling and energy recovery. Other sectors, such as agriculture and forestry, have also experienced a decline in production in recent years, and many unused crops or parts of trees in forests are rotting or infested with pests. In addition, in agriculture and forestry itself, many by-products are produced during the processing and extraction of the primary raw material. Municipal Waste Municipal waste consists of a wide range of different types of waste, most of which still have their use and their production is still increasing. For example, the production of municipal waste in our country Slovakia has increased in 10 years by 100 kg per capita to a value of up to 427 kg. The European average in waste production is 492 kg per capita, which means that in some more developed countries this amount is even higher. In our country, 54% of that waste ends up in landfills. Even if, 64% of municipal waste is biodegradable waste [6]. These include food waste, wood, paper, biodegradable textiles, but we cannot forget the sewage sludge, which is even collected in separate tanks. However, the largest usable share of biodegradable municipal waste suitable for incineration comes from the annual seasonal collection of grass, branches, leaves, and needles from our parks and gardens [7]. Much of this waste is also generated by urbanization and rethinking of new gardening styles. For example, China is increasing the area of its settlements by 6.1% every year, and for recent years this has brought it up to 12.7% of new cultivated areas that produce this green waste every year [8]. Agriculture In recent years, mankind has achieved a significant increase in the efficiency of agricultural production [9]. For this reason, we need smaller areas to feed the world’s population, and many areas become unused or crops grown on them lose their importance. Due to poor resource management, often grown food becomes waste that is at least incinerated. Worse fate awaits many secondary products of plants such as stalks and leaves of herbs and seed skins. These materials are no longer used due to the reduction in the number of herbivores reared for milk and due to changes in their diet for more nutritious crops using for faster meat rearing. Many of them either end up rotting in fields or landfills for bio-waste (see Fig. 3). Forestry Forest management also requires a reassessment of the use of many waste products. In the field of forestry, these are mainly leaves, bark, branches, and cuttings from logs. Many logging companies prefer to unnecessarily ignite these remnants in the woods to make space for new trees. The energy value of these raw materials is lost forever in this way. The increasing production of these waste parts of forest biomass is directly proportional to the consumption of construction wood used for the construction of houses and furniture and especially paper, which is used from writing literature up to packaging materials. However, with so much increasing

Closed Cycle of Biodegradable Wastes in Smart Cities

63

Fig. 3 Waste biomass: (a) rotting straw bale in the waterlogged field, (b) rotting tree branches in an illegal dump near human settlements

wood consumption, clean wood is still being burned in many places. Although the combustion of pure wood is carbon neutral, from a long-term point of view, it is better to produce products from the wood which can keep that carbon for a long time and burn mainly unused secondary waste products arising from its growth and processing [10]. By properly managing waste products from the timber industry, tree wastes parts in forests and urban parks, we can achieve more efficient use of renewable resources, reduce excessive pressure to cut down healthy trees and thus obtain a carbon-neutral source of clean energy suitable for coal replacement. These facts increase the amount of potential biodegradable waste and thus enable its use in the field of energy recovery (see Table 1).

2.2 Biodegradable Waste Processing The processing of such a diverse material as waste biomass is very demanding and requires different processing technologies that are determined by different combustion techniques. The most important factor influencing the use of waste biomass is its state. Most of the waste biomass is in the solid state, but we also know waste on a liquid and rarely gaseous basis. Gas Bio-wastes Gas-based bio-waste arises in the form of biogas, which is generated during anaerobic rotting of solid or liquid forms of waste biomass, most often in landfills or wastewater treatment plants. This technology of anaerobic digestion we can also use intentionally with agricultural waste biomass. The resulting biogas can become a perfect replacement for natural gas in gas heating plants using the right purification technology [12]. Liquid Bio-wastes Liquid bio-wastes contain a large amount of water and are therefore not suitable for direct incineration and are more often used for low-

64

M. Holubˇcík et al.

Table 1 Total usable potential of energy biomass in Slovakia [11] Type of biomass source Forest biomass Thin trees up to 7 cm Rough waste tree Wastes from handling of wood Fire wood Biomass from cuttings stumps and roots Wastes from the wood processing industry Agriculture biomass Cereal straw Rape and sunflower straw Orchard and vineyard waste Biogas Biodiesel Municipal waste Sludge from sewage treatment Mixed municipal waste

Mass (t/year) 2,168,030 250,740 76,200 214,390 323,900 14,300 23,500 1,265,000 542,136 272,700 161,300 50,400 52,236 5500 1,801,762 310,200 1,491,562

Energy (TJ/year) 24,631.18 2,383.05 724.00 2,220.69 3,079.81 138.58 223.25 15,861.80 7,799.90 3,861.00 2,223.30 528.60 972.50 214.50 22,090.88 2,528.00 19,562.88

temperature decomposition. However, there is still the possibility of drying them, whereby dry powder forms of fuels are formed, which can then be compressed into the form of pellets or briquettes. Solid Bio-wastes Solid biodegradable wastes have the largest volume and range of use, but also forms a wide variety of forms from powder through lumpy material to larger parts that form an inhomogeneous structure of shapes and sizes. For this reason, we need to grind and chop shapeless types of bio-waste to homogeneous matter. Properly shaped biomass can be fed to the combustion burners of thermal power plants in the required amount necessary to release a sufficient amount of thermal energy. By fine grinding, we are also able to replace conventionally used coal in thermal power plants, which is most often burned by powder combustion in fluidized beds [13]. However, solid biomass is usually burned in a moving grate incinerator, which does not require complex fuel preparation before combustion. In addition, we also know various specialized technologies of solid waste biomass combustion such as pyrolysis, bale straw combustion or combustion of lump compacted biomass in the form of pellets as briquettes which are more often used in small heat sources. Pyrolysis is a very old process and in the past it was mainly used to produce light gas which was the forerunner of today’s natural gas and even today is very popular. It is a special method of thermal decomposition of biomass without access to oxygen in addition to the heat released, it is also possible to produce gaseous and liquid products that can be secondary combusted or supply distribution networks as energy carriers [14].

Closed Cycle of Biodegradable Wastes in Smart Cities

65

3 Waste Biomass Combustion Systems We know several energy supply systems for cities that are based on two principles. The first historically older principle is decentralist energy distribution, which assumes that each household has its own heat source. The second more modern principle counts on only one heat source for the whole city, from which heat is distributed to individual households by means of distribution networks. In practice, any urban heat supply system is based on the most sensible compromise between the two principles.

3.1 More Decentralized Systems Less centralized systems are cheaper and more suitable for less centralized cities with more isolated neighborhoods. They are often used in the suburbs of large towns with terraced buildings. They are characterized by small heat sources from 5 to 100 kW designed for heating usually one but rarely several houses. In developed parts of the world, only a wealthier class of the population can afford such housing, which means that such neighborhoods are smaller and equipped with better combustion facilities. But in developing countries with a high population density and low-quality combustion devices, such as China, this method of heating is becoming a major problem because it significantly pollutes the environment with emissions [15]. This category also includes medium heat sources from 100 kW to 1 MW intended for heating blocks of flats in housing estates. Medium sources are used to heat one or more surrounding apartment buildings, while each apartment building contains several separate apartments. They are also used for heating public buildings with higher consumption where small resources are not enough. Solid fuel boilers for decentralized small heat sources are a standard and so far the most frequently used heat source. These combustion devices are able to burn mostly only large pieces of clean wood or briquettes which can also be produced from waste biomass. Briquettes are large-scale moldings of various shapes, such as a cylinder or block, which are pressed from finely ground sawdust on a briquetting line. In addition to solid fuel boilers, we also know more modern automatic pellet boilers, which are gradually replacing them and can also compete with an automatic natural gas boiler. Automatic pellet boilers are most often used as heat sources in medium systems because of their easy and carefree operation with fuel dosing. The pellets burned in them are produced by a similar process of pressing fine dust as briquettes, with the difference that they always have only a cylindrical shape of smaller dimensions and are pressed at once in larger quantities. Pellets can be pressed from pure wood or other alternative sources of biomass, and in addition to pellets, automatic boilers can also burn wood chips, which require much lower input energy costs for processing.

66

M. Holubˇcík et al.

3.2 More Centralized Systems More centralized systems are greener because they have significantly lower emissions and greater energy efficiency, whereas they can also cogenerate electric energy with thermal energy [16]. On the other hand, centralized systems have higher input costs, especially for combustion technology and distribution systems that only larger and more centralized cities can afford. Smaller cities are usually equipped with large heat sources with outputs from 1 to 10 MW which supply heat to all important urban settlements and other suburbs usually have their own decentralized heat sources. Larger cities need to be supplied with heat from super-large heat sources with outputs above 10 MW. Real megalopolis usually has more than one of such large sources that supplies heat to a centralized hot water or steam system, but even here, some peripheral parts are at least with terraced buildings solved by a decentralized system. In the production of such large heat outputs, the sizes of the combustion chambers are similar to the dimensions of family houses. Therefore, sometimes it is not necessary to process the fuel and many combustion lines with a movable grate are able to burn larger parts of biomass at once. This system also allows us to burn contaminated biomass containing ceramic inorganic materials difficult to crush. However, in smaller thermal power plants, the fuel is crushed into smaller parts in order to achieve optimal production of thermal energy, but on the other hand, this system is the least energy-intensive compared to specialized fuel preparation for small sources. On the other hand, large sources also commonly use the possibility of powder combustion of fine materials which requires grinding but not additional pressing. This method allows the combustion of only pure waste agricultural or forestry biomass in the form of grass or wood leaves [17].

4 Experimental Testing Methods For our research, we identified seven basic representative samples representing individual parts of tree and herbaceous vegetation, which include representatives of individual types of biodegradable waste from agriculture, forestry, and urban green areas. The identified properties of the investigated samples could serve as a tool for determining the convenient ratios of fuel mixtures that we would be able to obtain from waste biomass. Measurements could tactically reveal to us which of the samples has better properties and thus its use would pay off more than the use of other samples. For comparison, we used pure wood commercially used as a fuel and compared its properties with the properties of the examined samples (see Fig. 4).

Closed Cycle of Biodegradable Wastes in Smart Cities

67

Fig. 4 (1) Pure wood, (2) Corn cobs, (3) Lawn grass, (4) Field straw, (5) Tree leaves, (6) Mixed wood chips, (7) Conifer bark, (8) Conifer needles

4.1 Determination of Ash and Moisture Content Moisture content was determined according to STN EN 14774 together with the ash content on the thermograviometer LECO TGA 701 (see Fig. 5a). The test samples were gradually weighed into small crucibles weighing from 2.5 to 4 g. After starting the device, the combustion chamber first warmed up to the temperature 105 ± 2 ◦ C at which the water evaporated. The difference in the weight of the crucible showed the weight of the evaporated moisture. Then the device was heated to a temperature at which volatile combustibles were evaporated and volatile carbon burned immediately afterwards. The difference in the weight of the empty comparison crucible and the crucible with the burnt samples showed us the final weight of the ash.

4.2 Determination of Calorific Values Calorific value was determined on the calorimeter LECO AC 500 (see Fig. 5b) according to STN EN ISO 1716 [18]. Samples of waste biomass weighing about

68

M. Holubˇcík et al.

Fig. 5 Measuring instruments: (a) thermograviometer LECO TGA 701, (b) calorimeter LECO AC 500

1 g were placed in an iron crucible with which they were placed in a pressure vessel where we pressurized oxygen to 150 bar. The sample was ignited by an electrical short on a metal wire. The heat released during combustion was captured in the water in the calorimeter tank and the difference in water temperatures determined the Gross calorific value of the fuel. By subtracting the heat dissipated away by the evaporated moisture, we then calculated the value of the real calorific value from the equation:   Qi = Qs − 2453 (w + 9.H2 ) MJ.kg−1

(1)

where Qi is the calorific value (MJ kg−1 ), Qs is the gross calorific value (MJ kg−1 ), w is the moisture (kg kg−1 ), and H2 is the hydrogen content (kg kg−1 ) in the sample of the municipal waste. Three measurements for each sample were made and the resulting values were their average.

4.3 Determination of Chemical Content We used two LECO modules from the 628 [19] series to determine the chemical properties of the samples, one to determine the carbon, hydrogen, and nitrogen content and the other a separate module to determine the cheese content. Nitrogen, Carbon, and Hydrogen Module To determine the content of these three elements we used the combustion device LECO CHN628 (see Fig. 6). A bead with samples weighing approximately 1 g was placed in the combustion chamber and wrapped in aluminum foil. The furnace was heated to 1050 ◦ C under an atmosphere of pure oxygen when the entire contents of the sample burned and the contents of the individual elements were analyzed from the flue gases.

Closed Cycle of Biodegradable Wastes in Smart Cities

69

Fig. 6 LECO combustion device 628: (a) Module S and (b) Module CHN

Sulfur Module The determination of the sulfur content was determined using a 628 s module where we placed approximately 2.5 g of sample in a ceramic tray. A combustion chamber with also a high temperature above 1050 ◦ C in an atmosphere of pure oxygen also burned all the sample material. The resulting sulfur value was determined from the flue gas emissions.

5 Experimental Measurements Results 5.1 Energetic Properties The most basic property investigated for any combustion sample is the calorific value of the raw material. This value shows us how much heat we get from burning the material and thus how much raw material we need to burn to ensure the heat demand for the city. The second most basic property of the substance is its moisture content, which negatively reduces the calorific value. Gross calorific values and calorific values together with the moisture values of the samples will give us a basic understanding of the use of raw materials (see Table 2).

5.2 Chemical Properties The second important series of properties we focused on was the chemical content and the ash content of the samples. Ash is an inorganic component of the fuel which plays no role in combustion and forms ballast in the final product. The amount of ash in the raw material and its properties are responsible for a number of difficulties in maintaining a smooth combustion process. Too much ash causes suffocation of the fire and increases the cost of disposing of this waste product. Its low melting

70

M. Holubˇcík et al.

Table 2 Humidity and calorific values of the test samples

Material Pure wood Corn cobs Lawn grass Field straw Tree leaves Wood chips Conifer bark Conifer needles

(EU code) – (02 01 03) (20 02 01) (02 01 03) (20 02 01) (20 02 01) (03 01 01) (20 02 01)

Moisture (%) 7.12 5.67 7.26 7.67 6.35 9.54 3.18 6.52

Gross calorific value (MJ kg−1 ) 17.4398 17.3138 15.7185 17.7887 18.5324 16.2642 18.2679 20.3451

Calorific value (MJ kg−1 ) 15.9275 15.8126 14.3129 16.3135 17.2573 14.7468 16.8984 18.8716

point, especially with some types of biomass, significantly affects the combustion conditions, especially the combustion temperature in the combustion chamber. The elements of the CHNS group form a flammable component in the fuel. The most significant share is mostly carbon, which is the main component of most fuels we can get from nature, but its combustion produces an undesirable greenhouse gas, carbon dioxide, and sometimes even dangerous carbon monoxide. Elements such as nitrogen and sulfur are important to us in terms of the fact that their combustion produces dangerous nitrogen oxides and sulfur oxides, which mix with water vapor in the air and cause acid rains. The cleanest is the combustion of hydrogen, which is converted to harmless water vapor when burned (see Table 3).

6 Conclusions The results of our experiments showed that with certainty all the examined samples carry enough energy suitable for combustion. The lowest values were shown by the garden lawn due to the high content of the inorganic ash part, which is probably caused by excessive fertilization and frequent mowing. In addition, the chips did not fit well due to their high humidity and especially the contamination during processing. Other samples showed values about the same as pure wood, and some like pouring needles or bark even higher which could be due to the higher content of resins and more nutritious particles. From the point of view of moisture, the bark or needles of coniferous trees, which contain water-repellent resins directly in the cell structure, withstood the best. The other samples showed on average the same moisture contents, with the exception of wood chips, which contain a high percentage of shell bark, which protects the wood from drying out and therefore its drying time is considerably longer. The soft tissues of straw or grass leaves have the fastest drying time, but on the other hand they are also very absorbent.

Material Pure wood Corn cobs Lawn grass Field straw Tree leaves Wood chips Conifer bark Conifer needles

(EU code) – 02 01 03 20 02 01 02 01 03 20 02 01 20 02 01 03 01 01 20 02 01

Content of N (%) 0.06 0.17 1.89 0.76 0.09 0.05 0.77 1.17

Content of H (%) 6.05 6.17 5.56 5.83 5.07 5.57 5.85 5.95

Table 3 Humidity and calorific values of the test samples Content of C (%) 48.15 47.17 40.74 44.46 44.63 50.45 48.24 40.92

Content of S (%) 0.03 0.02 0.20 0.11 0.09 0.07 0.12 0.11

Ash content (%) 0.49 1.63 13.35 5.55 14.51 1.29 4.92 8.13

Closed Cycle of Biodegradable Wastes in Smart Cities 71

72

M. Holubˇcík et al.

In the content of the elements, we decided to evaluate each element separately. As for the hydrogen content, we noticed the best balance and did not notice any major differences, but the best place in this regard was the corn cobs. The carbon content was probably the most interesting because we noticed the biggest differences between the samples. Elements containing more woody matter, such as wood chips, bark, but surprisingly also corn husks, were characterized by its highest values. On the other hand, samples of soft structures such as needles, turf, straw or leaves were characterized by a lower value of bound carbon. On the contrary, the nitrogen content was the lowest in woody structures, especially in the case of pure wood and wood chips. Surprisingly, the nitrogen content in the leaves and corn cobs was very low, but here it could be assumed because it is more of a carbohydrate-tuned crop. We recorded relatively high values of nitrogen in the needles and the largest in the lawn, which is certainly caused by the excessive use of nitrogen fertilizers. The sulfur content is probably the most watched element in the field of emissions, and this category was also won by the lawn. The lowest value was reached by pure wood and corn husks and other samples reached average values. With such a large number of criteria, it is usually difficult to choose the most advantageous raw material, but we have certainly been able to determine which of the raw materials is the least advantageous in terms of emissions and calorific value. It is a lawn that I would not use for heating and it is questionable whether it would be suitable for fattening animals with so many artificial fertilizers. But even this raw material might have better properties if a more sensible fertilization approach was chosen. All other raw materials have shown sufficient energy properties suitable for heating plants and in some cases their emission footprint is still questionable, but as far as combustion in large sources is concerned, this has long been solved by means of separators and flue gas scrubbers. Acknowledgements This contribution has been created as part of the project VEGA 1/0233/19 “Construction modification of the burner for combustion of solid fuels in small heat sources” and KEGA 033ŽU-4/2018 “Heat sources and pollution of the environment” and APVV-17-0311 “Research and development of zero waste technology for the decomposition and selection of undesirable components from process gas generated by the gasifier”.

References 1. Mishra UC (2004) Environmental impact of coal industry and thermal power plants in India. J Environ Radioact 72:35–40. https://doi.org/10.1016/S0265-931X(03)00183-8 2. Karampinis E, Grammelis P et al (2014) Co-firing of biomass with coal in thermal power plants: technology schemes, impacts, and future perspectives. Wires Energy Environ 3:384– 399. https://doi.org/10.1002/wene.100 3. Rasmussen C, Vigsoe D (2005) Rethinking the waste hierarchy. Institut for Miljoevurdering, Copenhagen, p 118 4. Kammen MD, Sunter AD (2016) City-integrated renewable energy for urban sustainability. Science 352:922–928. https://doi.org/10.1126/science.aad9302

Closed Cycle of Biodegradable Wastes in Smart Cities

73

5. Zavodska A, Benesova L, Smith B, Morrissey A (2014) A comparison of biodegradable municipal waste (BMW) management strategies in Ireland and the Czech Republic and the lessons learned. Resour Conserv Recycl 92:136–144 6. Pobozna M (2019) Waste in the Slovak Republic 2018. Statistical Office SR, Bratislava, p 92 7. Crowe M, Nolan K; Company Biodegradable Municipal Waste Management in Europe (2002) European Environment Agency, p 123. http://waste.com.br/textos/ Biodegradable%20municipal%20waste%20management%20in%20Europe.pdf 8. Shi Y, Ge Y, Chang J, Shao H, Tang Y (2013) Garden waste biomass for renewable and sustainable energy production in China: potential, challenges and development. Renew Sust Energ Rev 22:432–437 9. Loehr CR (1974) Agricultural waste management problems, processes, approaches. Subsidiary of Harcourt Brace Jovanovich, Ithaca 10. Macfarlane WD (2009) Potential availability of urban wood biomass in Michigan: implications for energy production, carbon sequestration and sustainable forest management in the USA. Biomass Bioenergy 33:628–634 11. Cicmanec S (2008) Biogas - suitable supplement to natural gas. Slovgas 5:24–27 12. Hrobaj P (2000) Ecological aspects of combustion. Neografia, Martin 13. Ohman M, Nordin A et al (2000) Bed agglomeration characteristics during fluidized bed combustion of biomass fuels. Energy Fuel 14:169–178. https://doi.org/10.1021/ef990107b 14. Puy N, Murillo R et al (2011) Valorization of forestry waste by pyrolysis in an auger reactor. Waste Manag 31:1339–1349 15. Fang W, Song W, Liu L et al (2000) Characteristics of indoor and outdoor fine particles in heating period at urban, suburban, and rural sites in Harbin, China. Environ Sci Pollut Res 27:1825–1834 16. Caserini S, Livion S, Giugliano M, Grosso M, Rigamonti L (2010) LCA of domestic and centralized biomass combustion: the case of Lombardy (Italy). Biomass Bioenergy 34:474– 482 17. Obernberger I (1998) Decentralized biomass combustion: state of the art and future development. Biomass Bioenergy 14:33–56 18. STN EN ISO 1716, 2010: Reaction to fire tests for products. Determination of the gross heat of combustion 19. LECO Corporation Michigan, USA, 2019: 628 series elemental analysis by combustion. https:/ /www.leco.com/images/Products/elemental/628/628-SERIES_209-218.pdf

Automated People Counting in Public Transport Peter Pištek

, Simon Harvan, and Michal Valicek

Abstract This paper is elaborating on problems in public transport in the context of Smart Cities and Internet of Things (IoT). The means of public transport are commonly overcrowded and on the other hand some lines could be designed inefficiently. This affects passengers’ comfort and also the financial site of public transport companies. The field of counting people in public transport is specific with its variety and limitations regarding setup in vehicles, which we took into account while designing the embedded system. We propose a solution—an embedded system with an array of infrared sensors. Approach uses image processing means (gauss filter, sliding average, and thresholding) for object detection, followed by the detection of direction by correspondence of objects positions across images. We performed controlled experiments with one and four participants in different scenarios which were compared to other similar solutions. We have achieved satisfying results up to 95%. Keywords Counting · Public transport · Infrared sensors · Smart cities · Embedded systems

The original version of this chapter was previously published with incorrect chapter author surnames. A correction to this chapter is available at https://doi.org/10.1007/978-3-030-929688_13. P. Pištek () · S. Harvan Kempelen Institute of Intelligent Technologies, Bratislava, Slovakia e-mail: [email protected] M. Valicek FIIT, Slovak University of Technology in Bratislava, Bratislava, Slovakia © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023, corrected publication 2023 D. Cagáˇnová, N. Horˇnáková (eds.), Industry 4.0 Challenges in Smart Cities, EAI/Springer Innovations in Communication and Computing, https://doi.org/10.1007/978-3-030-92968-8_5

75

76

P. Pištek et al.

1 Introduction The goal of this paper is to create a method for counting people in public transport. This is now a discussed topic among researchers, who are focusing on smart cities. According to the United Nations organization, urban areas are going to take in 68% of the world population by 2050 [1]. However, cities cannot grow much bigger, city architects and engineers have to find solutions to optimize energy, fuel consumption, processes, and comfort of people. Overcrowded public transport has an effect on waiting time, dependability and in the result on final choice between public transport and its alternative [2]. When public transport companies ignore these facts, it can result in the decrease of demand and ultimately in discomfort of people [2]. Knowledge of how many people are in public transport vehicles can provide important data for optimization of routes and deployments of buses, trams, etc. This would result in savings and increase in customer satisfaction. Regarding smart city concepts, this information about flow of people can be used for further optimization of other components, like estimation of how many people are currently in the city center.

2 People Counting Environment of public transportation may be described as dynamic. We have to assume that light (sun, shadow, day and night), temperature conditions (weather, year seasons) will be constantly changing. In addition, we have to also assume nonhomogeneous conditions within the scanned area. Some of the systems were having difficulties with this phenomenon and we would like to address this problem in design and implementation. We need to consider that movement of the interior might be an obstacle in the estimation of the crowd inside. As we have seen in papers [3, 4], estimating crowds is sensitive to movement and can worsen estimation. Another obstacle we might be facing is that some passengers are sitting rearfacing and some facing to the front of a bus. This means that a camera or another sensor might not recognize a passenger as a passenger. In general, we can divide methods for people counting (with means of computers) into two types, according to how they estimate number of people: 1. Estimating the number of people in a group. 2. Counting every person separately.

Automated People Counting in Public Transport

77

2.1 Estimating Number of People in a Group There is an effective estimation method for big crowds in open air spaces, that can be done without the help of computers. It is called Jacob’s method [5], which is also used nowadays. Its inventor proposed several measures: for loose crowd (3m2 /person); tighter crowd (1.4m2 /person); and really tight crowd (0.8m2 /person). Overall area multiplied by those numbers is the estimation of a crowd. Similarly, the size of a crowd can be estimated from images of the crowd and a trained neural network [6]. The number of people can be estimated based on the physical properties of some variables [3, 7, 8]. In these papers, they describe how the radio waves propagate and estimate size of crowd based on a change from normal state. The disadvantage of these solutions is: they do not expect movement of people and thus they have worse accuracy. Results in our environment would be even worse, because of constant moving of objects. We are not aware of an effective method for counting grouped people indoors or placed in a random manner.

2.2 Counting Every Person Most used designs, which are still used are turnstiles, which are used in enclosed networks such as the subway. However, this would not be efficient at bus stops, because we would lose flexibility of moving bus stops. Another design, which is fairly easy to deploy in condition of having such rules, would be counting the number of people signing their travel ticket. This would not be possible in many cities, where person does not have to sign the ticket prior to journey. This design has flaws, because there are stowaways, children, people with long-term passes for public transport and they would not be counted. Counting every person in public transport can be divided into three separate categories as follows: 1. Counting directly by means of image processing. 2. Counting directly by means of non-image processing (sensors). 3. Counting indirectly by counting features, traits of humans.

Counting Directly by Means of Image Processing The image is captured by a camera and each image is analyzed. At first, it is necessary to extract the background in the image [9]. Then we need to identify a person’s face (usually the head, sometimes with the shoulders, possibly with the torso). After identifying each person, the direction is assessed by evaluating several images before and after specific time. Thanks to the familiar location of the camera, we can determine where the object moves [8, 9].

78

P. Pištek et al.

These methods have obvious drawbacks including the lighting conditions that have to be provided. For commercial solutions, there is a need for 300 lux (that is during the daylight-saving day inside) to work with the usual accuracy. From a technical point of view, they also have the disadvantage because of a difference in the performance depending on how the person passes (e.g., person without backpack and person carrying backpack). In addition, the critical disadvantage of image processing methods is the privacy problem. IoT is based on Internet access and an image-based sensor can cause discomfort to people who are being monitored. Although videos are not stored, there is always the risk of abuse [10]. When we are talking specifically about cameras, prices of such devices are usually several orders of magnitude higher than typical infrared or ultrasonic sensors and we can say the same about energy consumption.

Counting Directly by Means of Non-image Processing (Sensors) In order to overcome the problems with image-based methods, sensors became a viable option. In the study [11], for counting people in public transport two passive infrared sensors (PIR) were installed. PIR has the advantage of being able to work in an environment without light and works independently of the appearance of a person (e.g., height) by being based on infrared radiation from the human body. However, it is often a problem to detect more people concurrently going against each other [12]. It also has to be adapted to the thermal conditions in a dynamic environment, which is also a challenge when using this type of sensor. In other studies [4, 13] they introduced methods that used radio wave imaging. In Fig. 1, we can see how the impulse radio ultra-wideband (IR-UWB) sensors were set up, although this work had a satisfactory success rate of over 90%, but for our Fig. 1 Scheme of IR-UWB sensors in paper [13]

Automated People Counting in Public Transport

79

defined problem was not applicable. We will have to place sensors on the top of the doors, which is a direct consequence of the nature of the environment (e.g., bus). This solution had also its limitation, that people were counted only when walking separately [13]. In the later paper [4], they proposed estimation based on reflection of radio waves in the closed room. In other studies [3, 14], the problem was defined, such that people were not moving (e.g., the number of people sitting in the room). On the one side, there was a transmitter, and on the other side of the room was a receiver, depending on how the signal changed, the number of people in the room was estimated. Such a setting, as far as we know, has not been tested in public transport. However, we believe that such a setting would not provide desired results, because of the nature of public transportation and how people move or sit in, e.g., buses.

Counting Indirectly by Counting Attributes, Traits of Human Indirect human counting is based on counting other than human attributes. We might say that everything that does not belong to two previous two belongs to this category. Simply put, it is counting property that is correlating with the number of people present. Such counting can include counting based on the number of phones trying to connect to the Wi-Fi network [15]. These types of measurement are less accurate but can be implemented more quickly and cheaper thanks to the technologies already available [15]. Another indirect counting is based on sensing carbon dioxide concentration [15, 16]. Papers propose such a method, but we could not find any real results. It would be definitely affected by air-conditioning of vehicles or buildings. Sensors also suffer from delay in CO2 concentration increase following an increase in occupancy [17]. Ideas worth visiting again would be measurement, that would be based on the weight of the bus or its traction with the road.

2.3 Peripherals Passive Infrared Sensor Passive infrared sensor (PIR sensor) is based on an electromagnetic sensing element, which measures infrared radiation emitting from all objects with temperature above absolute zero in the field of the sensor. It is most commonly used in motion detection scenarios [18]. Term passive is used, because sensors based on this technology are using little energy for work [5, 6]. The sensor element pairs can be connected as opposed inputs to the differential amplifier. Continuous exposure to high energies can still saturate the sensor materials and cause the sensor not to record more information. At the same time, this arrangement minimizes interference in normal mode, allowing the device to

80

P. Pištek et al.

withstand the trigger due to close electric fields [19]. However, if the sensor is placed in a changing environment and part of the field of view is in a different environment (i.e.: outside the bus), it may have consequences for triggering the sensor even, if it should not. It is worth mentioning that we can arrange PIR sensors into arrays. They operate on the same principle as individually, but create an image, which we can then be processed [20, 21]. Most of the sensors are 8 × 8, which creates an image with 64 pixels. We use a relatively new sensor (MLX90640), which has 24 × 32 sensors, which creates an image with 768 pixels. This promises better results, than before. IR-UWB Impulse Radio Ultra-Wideband (IR-UWB) is a radio wave-based sensor. Like Wi-Fi or Bluetooth, it has a transmitter and receiver. If we want to determine the distance from the object, we have to determine the so-called Time of Flight (the measured time from sending the signal upon receipt). This technology is mostly used for positioning in buildings. Since it has a relatively high frequency, we can accurately (5–10 cm) calculate the distance of two devices [22]. In the paper [13], they propose a method for counting people with the help of this type of sensor. They place two IR-UWB sensors parallelly next to each other approximately 50 cm from each other. This setting had relatively good accuracy, but it had several bad traits. Whole system was very sensitive to deploying since both sensors had to be in the same angle with the floor approximately 60◦ . This was, of course, specific to this setting, but some of those poor traits were consequences of IR-UWB properties. Close settings of two IR-UWB were mutually canceled out. This meant a longer distance between those two sensors and that implied somebody else might enter the counting zone. Ultrasonic Sensor Ultrasonic sensor is a sound wave-based sensor for measuring distance to objects using the physical property of sound waves and reflection from hard surfaces. When the sensor picks up the echo, the microcontroller calculates the time between sending and receiving the signal to get the distance from an object. Mustapha et al. [23] proposed a method for detecting obstacles, which has very good accuracy using ultrasonic sensors, but generally (to our knowledge) it is not used to recognize separate objects (whether there are more obstacles). It might be a good sensor for getting the system out of sleep mode, because of its relatively low energy consumption. But soundwaves would be undesired in the presence of pets (especially dogs), since these sensors are working on 40 kHz frequency and this is within dogs hearing range (out of humans). Next disadvantage of this sensor is the inconsistency of waves propagation in some environments (e.g., foam on the surface of a liquid, change of atmosphere), which may cause measurement uncertainties.

Automated People Counting in Public Transport

81

3 Proposed Solution Based on the analysis, we propose a method of counting people (entrances and exits from vehicles). We need to place counting nodes above each entrance of a vehicle. Counting nodes are going to count entrances and exits, based on an array of passive infrared sensors. After pre-processing and object detection, we are going to detect the direction of the object (in our case, objects might also be in ease). After counting is finished, we will send data to the main node, which is not only responsible for counting, but also for sending data to the server. Before it is sent, we will get the GPS location of a bus and send it together with counting data (see Fig. 2).

3.1 Counting Node Design Circuit diagram can be seen in Fig. 3. We are using ESP8266 with a development board from manufacturer NodeMCU. ESP8266 has a large community of developers, which created and integrated specifics of this board into Arduino IDE. List of used components: • NodeMCU development board with ESP8266 with Wi-Fi chip on it. This development board is running at maximum clock frequency 160 MHz. • MLX90640BAB IR array 32 × 24 with field of view 55◦ × 35◦ . • HC-SR501 PIR motion sensor. • HC-05 bluetooth module. • SUP500F GPS module.

Fig. 2 Diagram showing simplified process of counting people

82

P. Pištek et al.

Fig. 3 Circuit diagram of main counting node

During initialization we will take background temperature for later processing optimization as is mentioned [24]. This will be important in the separation of objects, background and will help recognize background in various conditions (temperatures) [11]. In a vehicle there will be several counting nodes (one above each entry). Processing of data will be on the side of a counting node, which already has enough computation power partially taken by image processing (has to be evaluated by experiment). Data processing could be shifted to the server side, but there would be probably communication problems, because of big chunks of data or problems with signal strength (mobile or Wi-Fi). Each counting node will look the same except for the main counting node which will also have a GPS module (SUP500F on Fig. 3).

3.2 Pre-processing and Human Detection After initialization, an array of infrared sensors (MLX90640—32 × 24) provide data. This will create an image with 768 pixels, that will be preprocessed later. Pre-processing will consist of two steps as follows: 1. Smoothing. 2. Edge enhancement.

Automated People Counting in Public Transport

83

First smoothing, which is a common practice to reduce noise. Our approach uses either an unweighted sliding average, which is optimum for white noise removal [25] or a Gaussian filter which is widely used and is known for its good performance [26]. To prepare image to detect objects edge detection, we used method from Prewitt [27]. Human detection will be done by thresholding, we expect that threshold value based on initialized background temperature, but we also use a method of auto thresholding [28]. Areas with perimeter larger than a threshold will be considered human. Then we need features: • size of object, • temperature, • spatial position. These features are used in process of determining direction.

3.3 Determining Direction Determining direction is a demanding task, requiring analysis of more pictures. Based on paper [24], we expect that it is not possible to extract direction from a single picture, because of the resolution. We propose a method for choosing direction of movement as follows. From several images, we extract these features: • speed, • spatial position movement. The speed will be found out based on the method described in paper [29], where the camera has better resolution. Our solution needs only the speed to classify, whether a human is moving or just standing in sensors sensing area. Based on the second feature, we can estimate direction based on Y-axis movement and when speed is higher than a threshold, then a human is really exiting from vehicle.

3.4 Placing of Counting Node We are primarily focusing on placing nodes orthogonally to the door frame, in the middle of the door like in Fig. 4. Even though we were primarily focusing on this setup, we would also like to test a setup at the side of the door in the future.

84

P. Pištek et al.

Fig. 4 Placement counting node

Another important thing has to be considered. The field of view of sensor array, which is described in Fig. 5. In calculations of field of view, we took an average height of male in Slovakia, which is 179.4 cm [30]. With some space above the door frame we are expecting approximately 60 cm of space between the head and MLX90640 sensor. This gives us a field of view in the shape of isosceles trapezoid with base 55 cm sides 42 cm and top base 67 cm with 55◦ × 35◦ sensor. This is when the setup is at an ideal 17.5◦ angle to door.

3.5 Data Gathering The biggest challenge is to get data as fast as possible without regard for noise. Maximum possible refresh rate that we were able to get from MLX90640 was 16 Hz. At 32 Hz and more, our I2C frequency is too slow to acquire data from a sensor. I2C frequency is limited due to the software I2C interface of the development board.

Automated People Counting in Public Transport

85

Fig. 5 Field of view dimension in meters

Although, it is questionable, whether the noise would not be even more of a challenge than the number of frames per second. In Fig. 6, it is shown how the noise for each pixel increases with increased frequency. We can also see that the noise is bigger at the edges. After reading an output from calculation of data (subpages from the sensor) we could calculate the temperature, but it was not necessary for later image processing. This saved us approximately 20 ms at each frame. In samples of output shown in Fig. 7, we additionally calculated temperatures to show temperature precision.

86

P. Pištek et al.

Fig. 6 MLX90640BAB noise at each pixel and refresh rate at 4, 8, and 16 Hz [31]

Fig. 7 Samples of output from sensor MLX90640

We were able to read and calculate necessary calculations on average in 125 ms with setting refresh rate to 16 Hz. After reading the image we did an image preprocessing, which is described in next section.

3.6 Pre-processing We used methods Gaussian filter and sliding average to smooth noise [25]. Gaussian filter is combining neighborhoods of a pixel and multiplying it to the result. This is specific at the edges of pictures. When kernel comes out of image dimensions, there are specifically three methods to solve this: 1. Zero padding—adding 0 to enlarge image and then take only image dimensions

Automated People Counting in Public Transport

87

Fig. 8 Comparison of Gaussian and sliding average filter with different edge algorithms and different sizes of kernels. (a) Original image from sensor. (b) Gaussian filter 5 × 5 with edge mirror. (c) Gaussian filter 5 × 5 with edge value replication. (d) Gaussian filter 3 × 3 with edge value replication. (e) Sliding average with smooth width 3 and edge value replication. (f) Sliding average with smooth width 5 and edge value replication

2. Edge value replication—when the algorithm hits out of image, it takes a value of nearest edge 3. Mirror extension—when the algorithm hits out of image, it takes the value of mirror pixel We used the last two methods to produce images and then evaluated results. There is also a difference in the size of a neighborhood, that algorithm uses. For this purpose, we evaluated by visualizing the results (Fig. 8). We decided to use an edge mirror because it gives us the edge without a disruptive line of pixels. Regarding the Gaussian or sliding average, we had to do an experiment, which one would be more efficient, but we used the width of kernel 5, which gives us smoother results. We had to separate images without background data with foreground before thresholding, because it would result in false positives. We did an experiment with approximately 100 frames without anybody in the field of view and then 100 frames with a person at 50, 100, and 150 cm for 30 s each distance. This showed us that average intensity in frames with a person inside is higher and we could find a dividing line between frames with and without a person after using standard deviation and smoothing of average (see Fig. 9).

88

P. Pištek et al.

Fig. 9 Our comparison of standard deviation of intensities between frame with and without person after Gaussian filter Table 1 Comparison of Gaussian and sliding average filters performance

Event Entrance Exit

Ground truth 20 20

Sliding average 17 (85%) 16 (80%)

Gaussian filter 23 (85%) 20 (100%)

4 Experimental Results We needed to compare the accuracy of the described two filters (Gaussian filter and sliding) therefore, we asked one participant to walk through doors with sufficient time gap 20 times to each direction, while we were observing results. Results in Table 1 show us that the Gaussian filter is performing better in average. Although additional entrances were made by making noise larger and counting more persons in a few frames. From this point we used a Gaussian filter for smoothing the image in next experiments.

4.1 Scenarios One-Person Walking We performed an experiment for comparison with paper [24] with one man walking through doors in a controlled manner (Table 2). A participant was asked to walk through doors 150 times, 75 times out and 75 times into the room. The participant was dressed in t-shirt and walked with speed approximately 0.8 m/s. We mounted the sensor at 60 cm above the door frame (90 cm wide) and let the counting node train.

Automated People Counting in Public Transport Table 2 Our solution in one person walking scenario

89 Event Entrance Exit

Ground truth 75 75

Sliding average 77 (97%) 71 (94%)

Table 3 Special use cases of counting node Event Entrance Exit

Ground truth 20 20

With jacket 15 (75%) 12 (60%)

With stop beneath counting node 23 (85%) 20 (100%)

Table 4 Accuracy in four-people scenario Event Entrance Exit

Ground truth 40 40

Four in a queue 29 (72%) 29 (72%)

Two and two going in opposite directions 25 (60%) 24 (55%)

Special use cases (Table 3), that we tested, are walking with a jacket and hood overhead, and with stopping beneath the sensor. We wanted to examine the robustness of our algorithm and find edge scenarios and reactions of the algorithm to it. In the scenario with a hood over-head we completely covered the participant, so that there is no skin visible to the sensor (Fig. 10) and asked him to pass doors (90 cm) 40 times. As we expected accuracy dropped due to the sensor not being able to detect person in the field of view in several frames. However, the participant was asked to look down the whole time, like in Fig. 10. Second scenario with a stop beneath the counting node is simulation of bus entrance in which entrance is followed by quick stop. In this case, our solution has an average success rate 92.5%.

Four-People Scenario In scenarios with more people passing are two main cases. People going against each other or after each other. We asked participants to walk relaxed as we mentioned in section methodology and according to predefined scenarios. Accuracy (see Table 4) dropped due to more people in the field of view, which resulted in merged objects. Then we observe the accuracy drop in the case two against two, where people were asked to stay in boundaries.

4.2 Comparison with Other Methods After we measured controlled experiments, we could make the comparison with other methods dealing with counting people entrances and exits.

90

P. Pištek et al.

Fig. 10 Person with jacket and hood over-head

Table 5 Comparison of success rate in four in queue scenario

Event Entrance Exit

Our solution 72% 72%

Senti [11] 33% 50%

Table 6 Comparison of success rate with one-person scenario Event Entrance Exit

Ground truth 75 75

Our solution 77 (97%) 71 (94%)

Mohammadmoradi [24] 75 (100%) 72 (96%)

We are comparing our results with two other papers, which are counting people in public transport in Tables 5 and 6. Table 5 shows us the success rate in four in the queue scenario of our solution with Senti’s [11]. We compare Senti’s solution, where the experiment is described with the same method of four people passing through doors in a queue. Our solution counted with significantly better accuracy both entrances and exits.

Automated People Counting in Public Transport

91

Comparison is shown in Table 6, which shows us that the GridEYE solution (used in [24]) performs better, even though the total number of people it counted is less than in our solution. This is due to the fact that direction was wrongly estimated. Comparing results from one man walking with Mohammadmoradi [24], we have achieved similarly good results. Our solution with accuracy 95.5% and Mohammadmoradi’s solution with 98% accuracy. Total number of crossings was 150 in both experiments and our solution counted together 148 compared to 147, but their accuracy was better due to misclassified direction in our solution. This was due to the fact that our algorithm is not handling occlusions, when they happen in the middle of the path. In scenarios with stopping beneath the sensor, we dealt with the problem that the algorithm evaluated one crossing with stop as two crossings. We addressed this problem in our algorithm with setting all objects that already were in path as already counted in. However, these results showed us that there was a bug in our implementation that we have not found. However, when we looked into four people in a row (queue), we got significantly better results than Senti’s solution, which was based on two PIR sensors installed in row. Although they tested their algorithm on small samples, we have achieved fairly good results in more people scenarios. However, accuracy is dropping after there are more people introduced into the scene. When we looked at data generated in experiments, we found out that participants were merged together. That was caused partially because we were not able to control participants completely and they walked too close to each other and partially, because the sensor had a limited field of view, which we asked participants to fit in.

5 Conclusion The goal of this paper was to create efficient people counting algorithms. We analyzed the field of counting people widely with respect to people counting in public transport solutions. In analysis, we showed that this field is in an advanced stage and most state-of-the-art solutions have accuracy above 90% accuracy and many papers focus on additional parameters such as mounting flexibility, energy efficiency or combination of sensors to get even better results. We have chosen an array of infrared sensors, because of energy efficiency and privacy questions that may arise with using RGB cameras. While focusing on efficiency, we were thinking about use cases where application would be used and designed complex systems, that may be used in real environments. Designed embedded system (counting node) is placed above doors and counts entrances and exits from a vehicle. We propose an algorithm for counting people with counting nodes with regard to hardware used. Algorithm uses a smoothing filter and automatic threshold method that creates blobs in the image. We proposed marking objects with breadth-first search and then finding the trajectory of objects based on their positions. The

92

P. Pištek et al.

embedded system was then implemented together with a web application for the end user. Tests revealed us the constraints our system has (view of field, frequency of capturing frames). Our primary challenge is the frequency of capturing frames and following algorithm running time. We minimized calculations for each frame as much as it was possible. We completely omitted calculating temperatures for each frame and calculated only intensities of pixels, which are correlating with temperatures, so it did not have an impact on images. We were able to process one image below 150 ms together with the reading image. We tested our solution in controlled experiments. In the case of a one-person scenario, we obtained slightly worse results than a similar solution (95.5% compared to 98%) but in the scenario with four people in a queue we managed to get far better results compared to the existing approach (72% compared to 41.5%). Even though the results are promising there is a special case which is common in public transport: during rainy or cold weather there is possibility that people would be completely covered (without no visible skin). During this test, our results maintain 67.5% of accuracy in the one-person scenario which leaves the possibility to improve the proposed solution. Testing also revealed a problem in object tracking, that we were unable to locate, which would be possibly one of the factors, that is biasing performance of our solution. Furthermore, we would like to test our solution in different situations such as different numbers of people in controlled scenarios and also experiment with uncontrolled movement of people to compare our solution to other similar ones. Similarly, we could improve energy efficiency by putting the development board into standby mode during time in between stops.

References 1. UN Department of Public Information (2018) World urbanization prospects: the 2018 revision. UN Department of Public Information, New York 2. Tirachini A, Hensher DA, Rose JM (2013) Crowding in public transport systems: effects on users, operation and implications for the estimation of demand. Transp Res Part A 53:36–52 3. Cheng Y-K, Chang RY (2017) Device-free indoor people counting using Wi-Fi channel state information for Internet of Things. In: GLOBECOM 2017 - 2017 IEEE global communications conference. IEEE, Piscataway, pp 1–6 4. Yang X, Yin W, Zhang L (2017) People counting based on CNN using IR-UWB radar. In: 2017 IEEE/CIC international conference on communications in China (ICCC). IEEE, Piscataway, pp 1–5 5. McPhail C, McCarthy J (2004) Who counts and how: estimating the size of protests. Contexts 3(3):12–18 6. Wang C, Zhang H, Yang L, Liu S, Cao X (2015) Deep people counting in extremely dense crowds 7. Velipasalar S, Tian Y, Hampapur A (2006) Automatic counting of interacting people by using a single uncalibrated camera. In: IEEE international conference on multimedia and expo. IEEE, Piscataway, pp 1265–1268

Automated People Counting in Public Transport

93

8. Dittrich F, de Oliveira LES, Britto AS, Koerich AL (2017) People counting in crowded and outdoor scenes using a hybrid multi-camera approach 9. Song H, Sun S, Akhtar N, Zhang C, Li C, Mian A (2018) Benchmark data and method for real-time people counting in cluttered scenes using depth sensors 10. Grau A (2015) Can you trust your fridge? IEEE Spectr 52(3):50–56 11. Senti, Patrick (2011) Distributed people counting using a wireless sensor network 12. Wahl F, Milenkovic M, Amft O (2012) A distributed PIR-based approach for estimating people count in office environments. In: 2012 IEEE 15th international conference on computational science and engineering. IEEE, Piscataway, pp 640–647 13. Choi JW, Quan X, Cho SH (2018) Bi-directional passing people counting system based on IR-UWB radar sensors. IEEE Internet Things J 5(2):512–522 14. Yang X, Yin W, Li L, Zhang L (2018) Dense people counting using IR-UWB radar with a hybrid feature extraction method[-6pt]. IEEE Geosci Remote Sens Lett 2018:1–5 15. Kalikova J, Krcal J (2017) People counting by means of Wi-Fi. SCSP 2017:1–3 16. Li T, Fong S (2018) Counting passengers in public buses by sensing carbon dioxide concentration: system design and implementation 17. Meyn S, Surana A, Lin Y, Oggianu SM, Narayanan S, Frewen TA (2009) A sensor-utilitynetwork method for estimation of occupancy in buildings. In: Proceedings of the 48th IEEE conference on decision and control (CDC) held jointly with 2009 28th Chinese Control Conference. IEEE, Piscataway, pp 1494–1500 18. Yun J, Lee S-S (2014) Human movement detection and identification using pyroelectric infrared sensors. Sensors 14(5):8057–8081 19. Tsai CF, Young MS (2003) Pyroelectric infrared sensor-based thermometer for monitoring indoor objects. Rev Sci Instrum 74(12):5267–5273 20. Leech C, Raykov YP, Ozer E, Merrett GV (2017) Real-time room occupancy estimation with Bayesian machine learning using a single PIR sensor and microcontroller. In: SAS 2017 - 2017 IEEE sensors applications symposium, proceedings. IEEE, Piscataway, pp 1–6 21. Tyndall A, Cardell-Oliver R, Keating A (2016) Occupancy estimation using a LowPixel count thermal imager. IEEE Sensors J 16(10):3784–3791 22. Connell C (2015) What’s the difference between measuring location by UWB, WiFi, and Bluetooth? https://www.electronicdesign.com/communications/what-s-differencebetweenmeasuring-location-uwb-wi-fi-and-bluetooth. Accessed 15 January 2020 23. Mustapha B, Zayegh A, Begg RK (2013) Ultrasonic and infrared sensors performance in a wireless obstacle detection system. In: 2013 1st international conference on artificial intelligence, modelling and simulation. IEEE, Piscataway, pp 487–492 24. Mohammadmoradi H, Munir S, Gnawali O, Shelton C (2017) Measuring PeopleFlow through doorways using easy-to-install IR array sensors. In: 2017 13th international conference on distributed computing in sensor systems (DCOSS). IEEE, Piscataway, pp 35–43 25. O’Haver T, Pragmatic A (2018) Introduction to signal processing: with applications in scientific measurement, 2nd edn. CreateSpace Independent Publishing Platform, Scotts Valley 26. Karimi K, Dickson NG, Hamze F (2011) A performance comparison of CUDA and OpenCL 27. Prewitt JMS (1970) Object enhancement and extraction. In: Lipkin B, Rosenfeld A (eds) Picture processing and psychopictorics. Elsevier, London, pp 75–149 28. Ridler TW, Calvard S (1978) Picture thresholding using an iterative selection method. IEEE Trans Syst Man Cybern 8(8):630–632 29. Nagano A, Fujimoto M, Kudo S, Akaguma R (2017) An image-processing based technique to obtain instantaneous horizontal walking and running speed. Gait Posture 51:7–9 30. Evˇcíková L, Hamade J, Nováková J, Tatara M (2004) Growth and development trends in Slovak children and adolescents during the last 10 years. In: Životné podmienky a zdravie [Living conditions and health]. Editor, Bratislava 31. Melexis (2020) MLX90640 datasheet. Melexis, pp 1–60. https://www.melexis.com/en/ documents/documentation/datasheets/datasheet-mlx90640. Accessed 20 January 2020

Industry 4.0: From Smart Factories to Artificial Intelligence Václav Soukup

Abstract The subject of the study is the theoretic analysis and interpretation of the fourth industrial revolution referred to as Industry 4.0. Attention is primarily paid to smart factories and questions linked with the development of artificial intelligence. The main objective of the study is to describe the basic attributes of Industry 4.0 and typical development trends linked with the fourth industrial revolution. Partial objectives of the study include putting Industry 4.0 into a wider historical context and mapping security risks that modern information and digital technologies bring about. The study consists of four relatively independent parts in which various aspects of the researched topic are described and analysed from complementary points of view. The first part of the study applies an anthropological approach that allows the analysis of developmental changes in different types of manufacturing to make use of the terms “culture” and “sociocultural change.” The second part of the study maps the history of humanity from a historic perspective together with transformations of manufacturing technologies during the first, second, and third industrial revolution with emphasis on the use of energy resources. The third part of the study, discussing the fourth industrial revolution, presents basic visions of Industry 4.0 including its main aspiration, i.e. connecting the Internet of Things, Services, and People. The fourth part of the study deals with legislative and security risks resulting from the acceptance of new digital technologies and progress in the research of artificial intelligence. Keywords Industry 4.0 · Artificial intelligence · Industrial revolutions

V. Soukup () Vysoká škola mezinárodních a veˇrejných vztah˚u Praha, Prague, Czech Republic e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 D. Cagáˇnová, N. Horˇnáková (eds.), Industry 4.0 Challenges in Smart Cities, EAI/Springer Innovations in Communication and Computing, https://doi.org/10.1007/978-3-030-92968-8_6

95

96

V. Soukup

1 Culture as a Tool of the Human Race’s Adaptation to the External Environment Just as space and time are the basic attributes of matter, the basic attribute of the human race is culture. From the anthropological perspective, we can define culture as a suprabiologically conceived system of means and mechanisms through which people adapt to their external environment. From this perspective, the class of cultural phenomena consists of artefacts (material products of goal-directed human effort), sociocultural regulations (habits, manners, laws, taboos), and ideas (symbolic and cognitive systems). The first evidence of the genesis of human culture is Palaeolithic stone industries, old for several millions of years. Stone manufacture industries introduced in the life of humanity the phenomenon of cultural evolution. Since that moment, the existence of humanity did not exclusively depend on the whims of nature and the biological pressure of natural selection (biological adaptation), but also on cultural means (cultural adaptation). A defining feature of cultural evolution is constant acceleration whose speed and impact on nature is unparalleled in the universe we are aware of. The human ability to transfer artefacts, cultural technology, and findings cumulatively in time ensures continuity of human societies and allows for the constant growth of the culture. From this perspective, culture works as a non-genetic collective memory of humanity that emerges in the form of cultural heritage [1, 2]. What became the driving force of ever-accelerating development of culture in the history of humanity was an innovation that can be considered the source of endogenous sociocultural change. Processes of migration and diffusion that can be marked as mechanisms resulting in an exogeneous sociocultural change also had a great impact on the transformation of culture in space. During the cultural evolution, the humans’ relation to the nature was largely affected by the manner through which people obtained food, materials, and energy necessary for surviving from the ecosystem. The cultural adaptive strategy of humanity, at least from the perspective of evolution, developed gradually from gathering and hunting to economies based on pasturing and agriculture to modern industrial manufacture. A qualitative change in the way of adaptation to the environment was brought about by the Neolithic revolution, the emergence of first towns, and a shift of human societies to farming and pasturing thanks to which many more people could be sustained in a certain area. Subsequently, the development of farming and pasturing started changing the ecosystems of the Earth. Nature that was once relatively unaffected by humans became more and more integrated into cultural systems. The Neolithic period saw the beginnings of the transformation of the flora and fauna into the cultural landscape. During the subsequent cultural evolution, the range of human adaptation means and mechanisms kept on extending. A distinguishing feature of the cultural development in antiquity, the Middle Ages, and modern times was the constant growth in humans’ ability to gain control over the sources of raw and energy. Another significant source of the acceleration of human advancement was innovations in the sphere of spiritual culture. What

Industry 4.0: From Smart Factories to Artificial Intelligence

97

can be considered crucial from this perspective is the invention of printing press in the fifteenth century that allowed for better information transfer. Traditional economic systems got a completely new dimension of functioning by the capitalist method of manufacturing. Agricultural land meant no longer the fundamental source of livelihood, but it was replaced by industrial mass production linked with scientific and technical progress. Sadly, sophisticated cultural systems and modern technologies that were primarily intended to serve people, started turning more and more against the humans and nature in industrial societies.

2 Transformations of the World During Industrial Revolutions Mechanization of water energy and utilization of steam energy led to the first industrial revolution in Britain in the eighteenth century. Its integral part was growing urbanization, industrialization, and quantification of human labour that turned workers on the labour market into goods. From the perspective of the manufacturing method, this was a shift from the feudal society based on agricultural production to the manufacture and subsequent development of the steel and textile industry. The industrial revolution and industrialization of developed countries led to increased production and variety of the products manufactured in the nineteenth century and contributed to the expansion of Western civilization. Nevertheless, most workers in the developing industry, including women and children, were uneducated and suffer from poverty and inconsiderate exploitation. Between 1870 and 1914, the second industrial revolution pushed its way through in industrially developed countries (Britain, France, Germany, and USA). Its cause rooted in the field of obtaining and control of energy, in particular the invention of the engine with internal combustion. This new type of propelling vehicles and machines contributed not only to the development of industrialization, but also to an increased interest in sources of energy such as oil. A significant driving force stimulating industrial development at that time were also assembly lines allowing for mass production of industrial products or using electric power. What played an ever more important role was the development of communication media and transport infrastructure that allowed for transport of materials and goods. Dynamic growth of capitalistic economy led to an increase in the number of qualified engineers and technicians who, unlike most manual workers, mastered new technologies. Management also transformed significantly as it strove for scientific production management and systematic organization of workers’ labour. Among influential pioneers of the theory of “scientific management of production” before the First World War was Frederick Winslow Tylor who accentuated the creation of mechanisms allowing for optimum selection of workers following predefined criteria and their special training [3].

98

V. Soukup

The subsequent development leap in terms of quality and goods production is linked with the third industrial revolution whose beginnings date back to the 1960s and 1970s. Its sources were, again, scientific and technical innovations. This was a noticeable technological advancement in the sphere of computational science, electronics, information technologies and their use with the objective of manufacture automation [4]. At that period, communication became faster, the same applies to efficiency of business contacts. There was a constant increase in the efficiency and optimization of the manufacturing process. The number of new companies alongside with job openings kept growing. The management of manufacturing companies was becoming more aware of the need for investing funds in professional education of their employees and thus improve their specific knowledge and skills necessary for delivering high-quality work in their jobs. Research showed that educated employees accepted innovations more flexibly and they were ready to implement new technologies into the manufacturing process more efficiently [5]. In 2011 to 2013, scientists noticed that countries developed in terms of technology and information started entering the fourth industrial revolution referred to as Industry 4.0. The essentials of the revolution lie in introducing and utilizing Cyber-Physical Systems, i.e. automated systems that replace monotype tasks and monotonous occupations performed by people. According to the current prognosis, the focal point of the fourth industrial revolution will be “digital economy” through which the world of different things, processes, and phenomena will be remote controlled while it will remain in possible and actual integration and interaction. The three previous revolutions were product of human invention and innovation that led to radical changes in the sphere of manufacture, distribution, and consumption of material and immaterial assets. The source of technological progress was first steam used for propelling manufacturing machines, then electric power harnessed in mass production and eventually use of electronic, cybernetic, and computational systems for industrial production that became more and more integrated. The result of such innovations was also a vast process of a sociocultural change that led to society transformation and related cultural systems. The source of these development changes was technological progress. However, it did not affect merely the sphere of economic and manufacturing system, but it also reflected in other sociocultural systems such as politics, education, law, and ideologies.

3 Smart Factories at the Heart of Industry 4.0 The fourth industrial revolution is principally linked with performance of work activities through the Internet and use of new communication, information and cyber technologies in manufacture. The driving force of Industry 4.0 is interlinking the Internet of Things, Internet of Services, and Internet of People). Its main vision is that once it is transformed into a fully integrated, automated and internally interactive manufacturing environment whose basic units will be digitally controlled

Industry 4.0: From Smart Factories to Artificial Intelligence

99

automated manufacturing systems. Computer integration of engineering processes will encompass a comprehensive interlinking of the manufacturing chain from defining and designing a product, development of its desired parameters, production, testing, and quality control to distribution. The technological “heart” of the entire system will be new global networks utilizing interconnectedness of manufacturing units into Cyber-Physical Systems (CPSs). Such systems will be the basis of the structure of autonomous “smart factories” that will be able to conduct information exchange, process empiric data, and respond adequately to the changing market environment. Smart factories will be able to use, through the Internet, desired information databases, verify validity of manufacturing steps in relation to the changing market, and optimize relations between manufacturers, suppliers, and customers. Integration of various internal components and the possibility of their new configuration will allow smart factories to use the flexible manufacturing system in an optimum and efficient way [6]. The result of their activities will not be just “smart products”, but also new value chains, transformation of human resources, innovated business models, and increased competitiveness on the national and global level. It is likely that Industry 4.0 could greatly contribute to solving global environmental problems that were established by traditional capitalistic manufacturing systems. The fact that there are still more and more people living on the Earth, which results in the necessity to produce more and more food, is looming large in the existence of humanity and the effort to sustain constant economic growth. Furthermore, in order to maintain economic growth, it is necessary to cumulate more and more raw materials and energy. The result is a global environmental crisis that affects negatively the condition of hydrosphere, pedosphere, atmosphere, and biosphere. One of the consequences of human activities is, for instance, mass extinction of animal and plant species and a decrease in biodiversity. However, Industry 4.0 could contribute to optimization of raw material exploitation, reducing energetic and material demandingness of production, and positive impact on demographic growth. The key to solving global problems of mankind will be the analysis of extensive data sets (big data), introduction of autonomous robots in manufacture, computer simulations, semantic structures, cloud and use of digital date in the human–machine, human–human and human–nature relations. Industry 4.0 and its technologies it generates will, for instance, have a great positive impact on strategy of energy renewable resources, construction of environmentally friendly “intelligent buildings” and expansion of electromobility. Building “smart cities”, in addition to optimization of urban infrastructures and increasing efficiency of city transport, will contribute to reduction of fuel consumption of vehicles and a decrease in undesirable emissions. Implementation of Industry 4.0 will also include application of the strategy of increasing environmental-friendliness throughout various manufacturing industries. The basic precondition of successful functioning of Industry 4.0 will be horizontal and vertical computer integration of processes of planning, development, production, testing, and launching final products in the consumption sphere. On the level of horizontal integration of all sub-systems this will concern the reception of

100

V. Soukup

an order, implementation of production, and subsequent distribution of the product to the end customer. On the level of vertical integration of all sub-systems, all levels of manufacture, from the lowest degree of automated management of physical processes to management of the production section to human resources planning and ERP systems (Enterprise Resource Planning) will be structurally interconnected. This requires the creation and implementation of a new software environment that will allow for cooperation and system interconnection of modular cyber-physical systems. It will also be necessary to develop and implement platforms such as SOA (Service Oriented Architectures) that will ensure interaction between registered subsystems together with the integration of manufacturing units in different types of companies. Interaction between autonomous manufacturing and distribution units will also include efficient multidimensional communication based on negotiation and flexible cooperation. An important attribute of Industry 4.0 will be the nascency of “the Internet of things” where elements and complexes of the physical world will be interconnected through the Internet and information technologies. The key role will be played by software modules that represent physical elements actually existing in a virtual environment. In this environment, they will solve given tasks and coordinate their activity and use services that they will be provided to each other or that they evoke by their existence. The Internet of things and its software modules are closely linked with the Internet of Services to whose needs and requirements they respond interactively in the virtual space. Parallelly to the Internet of things and the Internet of services, there is also the Internet of people to cover needs for mobile communication (using through speech or visual or haptic communication). This will have special interfaces and will primarily serve people and robots with artificial intelligence. What contributed greatly to the creation and dynamic development of Industry 4.0 was acceptance of new technologies in the field of information transfer and processing, IT and digital systems, artificial intelligence, new materials, and biotechnologies. An important role was also played by the growth of cybernetics, robotics, mechanics, and system sciences and the creation of software modules allowing for interlinking the real and virtual world. The human factor was not negligible either. People are still considered part of the manufacturing process, even though some of them are not personally present in a manufacturing plant. What matters is that new technologies gradually created new models of socio-economic behaviour of people in their relation to communication, IT, and information systems. Industry 4.0 will pose new requirements concerning professional roles of implementation of new technologies and innovation of manufacturing processes is essential, so employees should perceive it as update of their job content [7]. It is likely that more emphasis will be laid on the growth of employees’ qualification, innovativeness, adaptability, autonomous decision making, ability of system, and interdisciplinary thinking. We can expect global transformation of the structure and job content of most occupations, changes in the labour market, and creation of new principles of work arrangement. It can be presumed that growth of digitization and automation will allow people work much more from home, which will affect their way of life as well as forms of family coexistence. New manufacturing technologies

Industry 4.0: From Smart Factories to Artificial Intelligence

101

will contribute to elimination of monotonous, physically demanding labour and will create new space for employees’ personal development. It is apparent that Industry 4.0 will bring along an end to many jobs linked with hard labour or routine administration work, thus decreasing opportunities for little qualified workforce on the labour market. At the same time, it will lead to new job openings, particularly in the field of work with databases and information systems and/or job activities linked with data set processing and protection. The market will value more and more experts in robotics, biotechnologies, IT, and communication technologies. Emphasis will be put on employees’ ability to combine traditional technical education with creative and innovative thinking and software skills. Desirable professional and personal characteristics of employees will be four basic areas of job competences. The ability to do the job autonomously, on the basis of specific impulses (personal competences), ability to communicate efficiently and collaborate with other employees and working teams (social competences), ability to transfer innovation into practice (implementation competences), and ability to apply specific knowledge and skills during the job performance (expert competences). Growing acceleration in the field of technological innovations is likely to cause that professional knowledge and skills will become quickly obsolete or they will be radically innovated and modified. Therefore, a requalification system enabling to retrain employees promptly will become an important part of the traditional education. Employment policy will require flexibility, support of desirable workforce, and creation of new job openings. It is probable there will be a shift in the current structural composition of study fields from humanitarian to exact technology subjects that will play a crucial role in building the new economic structure. There will also be an apparent shift to interdisciplinary oriented technological fields that will be a “bridge” between exact sciences and humanitarian fields. Quality and creatively conceived education will be able to respond to creation of new occupation will be preferred. In line alongside with the growth of the importance of digital technologies, it will be necessary to implement them into education. Digital technologies should become an educational tool allowing the formation of functional professional, social, and communication networks. More emphasis will probably be also put on the study of mathematics. An absolutely necessary innovation in the education system will certainly be accentuation of a comprehensive, interdisciplinary, and systemic approach that will allow for integration of findings from different scientific disciplines. Equally important is the imperative of a change in the “way of thinking” in relation to the changing manufacture systems and use of new information technologies. Education should thus primarily focus on acquainting students with how cyber-physical systems work. Last but not least, it is also necessary to improve knowledge and skills in the sphere of digital system security. As Industry 4.0 progresses, we will witness the birth of an education model enabling production of creative and flexible graduates capable of critical thinking, work with information, application of mathematical skills, autonomous decision making, and problem solving. We can also expect a change in the attitude towards teachers who play a crucial role in the education system. It will be necessary to improve their status in terms of their financial reward

102

V. Soukup

as well as in terms of their knowledge and teaching skills. Not only should teachers be good educationalists, but—more importantly—top class experts who can provide students with top notch education.

4 Humans in the Shadow of Artificial Intelligence: Risks of Industry 4.0 Before Industry 4.0 can fulfil its visions, it is necessary to carry out extensive research in the sphere of cybernetics and artificial intelligence. This is because all-embracing integration of all engineering processes requires first creation of manufacturing systems and interactive services that will be capable of self-learning, self-diagnose, self-repair and—on the basis of the information received—also optimize their activities throughout interactions with their environment. However, high autonomy and independence of such systems causes many concerns about their implementation into practice including security and the legislative status of artificial intelligence. Equally important is research in the sphere of virtual space where software modules of the Internet of things and Internet of services will operate. It should be noted that “Industry 4.0” is not a synonym for manufacture digitization or connecting machines to the Internet. The essence of Industry 4.0 is the use of artificial intelligence and new information technologies for the purpose of integration of and interaction between once independent manufacture systems and the relevant non-manufacturing activities and services. Structural changes of industry, economy, and other sociocultural systems evoked by Industry 4.0 will require a series of legislative adjustments. It will be necessary to adjust the current legislation framework as well as creation of the new one that will flexibly reflect the ongoing changes caused by human invention and innovation (endogenous sociocultural change) and changes caused by migration and cultural diffusion (exogenous change). It will also be necessary to take legal and legislative measures, especially in the field of industrial policy, labour market, science, research and digital agenda. In addition to normative measures such as audits, certifications, and monitoring of security it will be necessary to create and implement into practice legislation and legal measures that can be used in digital practice as well as when registering new sociocultural changes. Special attention needs to be paid to digital data protection as they represent an important commodity and require specific legislative and regulatory measures. Security measures have to be implemented comprehensively and systematically on all levels where digital data are stored and used. It is an extensive vertical security chain; it includes the lowest degrees of data and communication security, medium levels ensure infrastructural reliability and security of larger datasets and the highest degrees of digital data protection on the global level of entire companies. Companies should also operate in a virtual space that is reliably secure in all aspects and monitor on a constant basis possible threats of their datasets and have in place efficient digital and legal

Industry 4.0: From Smart Factories to Artificial Intelligence

103

protection of their intellectual property. Attention should also be paid to securing all functions of automated manufacturing systems. All of this requires—in addition to using adequate security technologies—also a clearly defined legislative and legal framework. The Internet also presents many security issues, which is given by the fact that it is a free and unregulated phenomenon “sui generis”. The Internet, in its substance, does not respect sovereignty of counties, ignores their boundaries as well as human privacy. In other words, despite its countless unquestionable positives it also poses a security threat [8]. The discussion on the topic of security also stirs development in the field of biotechnologies and research of artificial intelligence that can either be misused or get out of our control. Artificial intelligence can make autonomous decisions (not just merely respect predefined instructions) on the basis of data and information from the external environment that it will process. Furthermore, one of the attributes of artificial intelligence is its “nontransparency” that is sometimes referred to as the black box. We know of cases when it was impossible to trace down the reason of a particular decision made by artificial intelligence when assessing behaviour of thinking machines. Intelligence can be generally defined as the ability to solve problems and achieve complexly the set objectives [8, 9]. In this context it should be noted that humanity has not yet been able to create the so-called artificial general intelligence that would be able to behave and think like humans. There is no coincidence that scientists keep discussing about when an intelligent machine to such a degree can be established that it can solve different situations the same way as a human being would. That is why we consider it helpful to distinguish between “artificial general intelligence” and “narrow artificial intelligence”. Only humans now have the general intelligence that allows for achieving any chosen. Narrow intelligence that allows for achieving only certain or partial objectives is part of machines artificially made by humans. Artificial intelligence can also be viewed as software with a certain degree of autonomy that is able to generate its own decisions on the basis of obtained data. Scientific progress in this research area is testified by the fact that a “thinking machine” broke the famous Turing test in 2014: it managed to fool a person on the level of narrow artificial intelligence during an experiment so that the human was not able to detect that there was a computer on the other end of the conversation, but the human thought he was talking to a computer. At present, progress in the research of artificial intelligence is linked with the use of neuron networks and development of machine learning. There are algorithms that enable machines to learn similarly as people and animals do. Through learning and gaining new experience, such artificial manmade systems can respond more and more adequately to new situations. Naturally, we can question the ability of artificial intelligence to acquire such attributes of humanity such as will, empathy, self-awareness, and human emotions. Even though artificial intelligence does not fall in the same “human being” category, it masters problem solving, automated theorem proving, intelligent retrieval from databases, representation of knowledge engineering and expert consulting systems, pattern recognition, perception problems, natural language understanding, automated planning, combinatorial and scheduling problems and robotics and it keeps getting better at it. Naturally, all of these skills can be

104

V. Soukup

harnessed by trends linked with expansion of Industry 4.0. It cannot be doubted that “smart” machines, robots, and new generations of artificial intelligence will play more and more important role in the years to come not only in the sphere of manufacture and industry, but also in health care and services. Now we already have “intelligent” self-driving cars, robotic vacuum cleaners, medical diagnostic devices, and Google web search engine. Certain risk that is to enable creation of “general artificial intelligence” in 2025, according to some scientists, lies in that machines can achieve a higher level of intelligence than people. Then people are in danger of being looked down on to by intelligent machines with the same contempt that we now look at animals who we consider less intelligent. Who knows, maybe Stephen Hawking was right when he suggested that development in the field of artificial intelligence may give the upper hand to intelligent machines to such a degree that they overpass abilities of the human brain, improve their intelligence manifold times and will be in a situation when they will decide about the future of humanity. It is apparent that the fourth industrial revolution will bring about changes in the field of economic, social, political, and cultural capital. The question is whether people will really be happy and how the quality of human life will change under the determining influence of Industry 4.0. Happiness does not depend merely on civilization advancement, economic prosperity, and affluence. It is a well-known fact that higher quality of life and degree of happiness is manifested by people of less developed countries, whereas in members of developed Western countries, more and more people suffer from stress and boredom and they find the feeling of happiness in regular use of psychopharmaceutical substances. The question is whether people will or will not be more and more driven by algorithms under the influence of the fourth industrial revolution, just like machines [8]. Last but not least, what played an important part in human lives and the history of humanity as such, were meanings that people attach to the world. People in the past societies were always caught in cobwebs of meanings that they made themselves [10]. Just as ideology systems such as magic, religion, and science controlled human lives in the past, we are likely to create a new model of faith and a virtual world through modern biotechnologies and computer algorithms. However, this comes with the risk that they will have much more control over our lives than ever before. Thanks to new modern technologies, humans now feel to be not only “the selected great ape species”, but almost gods. Only future holds the answer to the question whether the existence of the fourth industrial revolution will bring us genuine happiness and authentic way of life. Acknowledgements This research was supported and funded by APVV-17-0656 titled Transformation of Paradigm in Management of Organizations in the Context of Industry 4.0

References 1. Soukup V (2011) Antropologie: teorie cˇ lovˇeka a kultury. Karolinum, Praha 2. Soukup V (2015) Prehistorie rodu Homo. Karolinum, Praha

Industry 4.0: From Smart Factories to Artificial Intelligence

105

3. Taylor FW (1911) The principles of scientific management. Harper & Brothers, New York 4. Wu J, Guo S, Huang H, Liu W, Xiang Y (2018) Information and communications technologies for sustainable development goals: state-of-the-art, needs and perspectives. IEEE Commun Surv Tutorials 20(3):2389–2406 5. Bartel AP, Lichtenberg FR (1987) The comparative advantage of educated workers in implementing new technology. Rev Econ Stat 69(1):1–11 6. Wang S, Wan J, Li D, Zhang C (2016) Implementing smart factory of industrie 4.0: an outlook. Int J Distrib Sens Netw 4:1–10 7. Morris MG, Venkatesh V, Ackerman PL (2005) Gender and age differences in employee decisions about new technology: an extension to the theory of planned behavior. IEEE Trans Eng Manag 52(1):69–84 8. Harari Y (2017) Homo Deus: Struˇcné dˇejiny zítˇrka. Leda 9. Tegmark M (2017) Life 3.0: being human in the age of artificial intelligence. Penguin Books, New York 10. Geertz C (1973) The interpretation of cultures. Basic, New York

Sustainable Urban Mobility–Multimodality as a Chance for Greener Cities: Evidence from Slovakia Mária Holotová , L’udmila Nagyová ˇ and Dagmar Cagánová

, Tomáš Holota

,

Abstract The importance of sustainable mobility has also been confirmed this year by the global pandemic of COVID-19, which has reduced the mobility of the population, thereby it has significantly reduced the level of dust, noise, and air pollution from the traffic in most European countries, including the Slovak Republic. App-based and shared-ride services have become highly popular and offer a level of convenience unseen before in the urban mobility systems all over the world. Individual car transport dominates at the expense of sustainable modes of transport in most Slovak cities. The city of Nitra is no exception, as the high number of trips during the peak-hours often leads to severe traffic congestion. One way to contribute to better condition is multimodality that allows urban residents to choose from a range of alternative travel choices. The aim of this paper is to assess the possibilities of multimodality in the context of short distance moves in the city of Nitra as well as to analyse how the change of mode choice variability affects the urban mobility behaviour. For the purpose of meet the objectives of this paper, a marketing research was conducted. The research findings show that the travel behaviour of Nitra’s citizens does not show the elements of sustainable urban mobility as the current infrastructure and overall opportunities are limited. Our findings point to significant differences in attitudes of residents from different urban areas. Keywords Mobility · Multimodal behaviour · Public transport · Sustainable transportation

M. Holotová () · L’. Nagyová · T. Holota Slovak University of Agriculture in Nitra, Nitra, Slovak Republic e-mail: [email protected]; [email protected]; [email protected] D. Cagáˇnová Slovak University of Technology in Bratislava, Slovak e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 D. Cagáˇnová, N. Horˇnáková (eds.), Industry 4.0 Challenges in Smart Cities, EAI/Springer Innovations in Communication and Computing, https://doi.org/10.1007/978-3-030-92968-8_7

107

108

M. Holotová et al.

1 Introduction The society is facing an urban renaissance. A growing share of the world’s population is located mainly in urban areas and it is assumed to reach nearly 70% by 2050 [1].With more than 50% of world population living in urban areas today, it is imperative to support heedful planning of city infrastructure, especially urban transportation [2]. Slovakia, like other countries in Central and Eastern Europe, feels the negative consequences of the expansion of individual car transport over the past 25 years. Patterns of traditional urban mobility are changing and require to be modified in order to meet future social requirements with regard to efficient and comfortable transport as well as reduced air pollution and noise [3]. The issue of mobility is very extensive, broad-spectrum, time-consuming, financially and professionally demanding [4]. People all around the world use private cars to travel to work. Most of these commuter trips are single-occupant vehicle trips. In the USA, for example, singleoccupant trips represent approximately 77% of all commuter trips [5]; similar percentages are found in Europe [6]. The situation in Slovakia and its urban locations is also critical, and transport does not show the elements of sustainability (Fig. 1). In addition to public transport, individual car transport, which records year-on-year increases, also contributes to satisfying transport requirements. In the observed period of 2005–2018, the increase in persons transported by individual transport represented 9.9%. The long-term unfavourable development in the modal split in favour of roads, in particular, individual (non-public) transport represents one of the most crucial issues of the Slovak transport sector which needs to be solved in coming years.

Fig. 1 Development of individual road transport [7]

Sustainable Urban Mobility–Multimodality as a Chance for Greener Cities:. . .

109

Similar problems can also be observed in other Central European countries that have almost the same economic development. A sustainable integrated multimodal transport system that meets the society’s economic, social and environmental needs should be developed in every urban transport system [8]. Multimodality has become a widely used buzzword in many disciplines. In transport, multimodality describes a combined transport system with the use of many different means of locomotion [9]. The idea of multimodal public transport is increasingly gaining traction and going forward in many researches [10]. It has the potential to contribute to cleaner, smarter, and more sustainable transport, shifting the mobility of passengers and goods away from roads, making the optimal use of infrastructure and reducing costs [11]. Experts, different groups of stakeholders and local authorities are more often promoting the situational combination of various transport modes as a significant key to achieving more sustainable urban mobility in the future [12]. All these efforts lead to the opportunity that the strengths of the green modes and newly emerging mobility services, such as car sharing and bike sharing, can be linked to provide a smart alternative to the private cars, what may then contribute to a reduction in the car usage [13]. The new mobility services are well developed in various metropolitan regions and are increasingly integrated in the existing public transport structures, e.g. by multimodal booking and information options [14, 15]. Despite the impacts of multimodality on the individual and household level, several studies consider the attributes of the wider urban environment being relevant. In objective terms, this mostly involves referring to a city size, distances, and urban form [13]. Nobis determined a positive relationship between a municipality population size and multimodal behaviour, given that public transport is involved in the mode combination [16]. To minimize the negative impacts of car travel, local governments promote the use of public transport. Unfortunately, many suburban and rural areas are not sufficiently served as they lack the population density to justify the public transit, in other words, public transit is not economically viable [17]. Public transport multimodality support is the only way to tackle transport in a sustainable way, it is a solution that is more efficient, cheaper, more environmentally friendly and contributes to the quality of life in the city. Transport in an urban area based primarily on individual road transport cannot work well. This use of transport space is inefficient, and the space requirements of road transport are almost impossible to satisfy [9]. The study of how people jointly use different travel means is one of the key issues in the contemporary transport research. However, measuring multimodality behaviours presents some intricacies that deserve more attention in order to come up with an instrument that is effective both on a modelling and on a policy viewpoint [18]. Towns can enhance and develop urban mobility plans by taking into consideration all attributes of sustainable urban strategy. Well-planned and efficient urban transport strategy combining efficient services of public transport, pricing schemes, strategic accessibility, land-use planning, passengers’ movement strategies, and infrastructure for green transport modes are the foundation for urban prosperity. However, because of many reasons, including diverging interests, at the moment there is no consensus in politics and society regarding what these solutions

110

M. Holotová et al.

should finally look like [3]. Finally, it is necessary to mention the need to ensure that the paradigms of smart urban mobility and sustainable urban mobility are aligned. To a significant extent, this is about bringing technological and social considerations closer together and ensuring that due importance is given to both [19].

2 Material and Methodology The main aim of our contribution is to assess the possibilities of the use of multimodality in the context of short distance moves in the City of Nitra. Based on the questionnaire survey conducted among the inhabitants of the selected city, we analysed how the change of mode choice variability affects the urban mobility behaviour. More specifically, we asked if the dominance of a certain mode in the city makes multimodal mobility behaviour more likely, in average. Our preliminary study showed that integrating different sharing systems in combination with the urban public transport can significantly enhance the urban mobility. For supporting our goals, we processed different kind of information. For theoretical background, secondary data were processed. Primary data used in presented paper were obtained from marketing research, which was conducted in the period March–May 2020. The questionnaire was prepared in two forms—online and printed. Online ones were processed in Google Forms and were available on social network (on Facebook within the groups that bring together the residents of individual urban areas) and sent by electronic mails. Approximately 7% of the total number of completed questionnaires consisted of questionnaires completed in printed form at personal meetings. Statistical data processing was performed via XLStat. We used the ChiSquare test of independence and Correspondence analysis, as a deeper analysis was necessary. Lastly, 1840 citizens of Nitra were involved; as it is shown in Table 1, where individual characteristics of respondents are listed (the age ranges are limited according to the Pew Research Center methodology [20]).

3 Results and Discussion 3.1 Why a Sustainable Urban Mobility Project is Necessary in the City of Nitra There are many different approaches to urban mobility planning in Europe, with countries such as the United Kingdom or France at the forefront of this process. The policy of urban mobility, including planning, is primarily the responsibility of local public authorities that have been struggling with a lack of funding, insufficient legislation or the absence of professional staff for a long time. However, the current state of mobility in our cities is caused mainly by the lack of interest to look for

Sustainable Urban Mobility–Multimodality as a Chance for Greener Cities:. . .

111

Table 1 Socio-demographic characteristics of respondents Socio-demographic variables Gender Age group

Economic activity of respondents

Urban area

Categories Female Male 18–23 24–39 40–55 56–74 Employed Unemployed Student Maternity leave Retiree ˇ Cermᡠn Dolné a Horné Krškany Drážovce Chrenová Janíkovce Klokoˇcina Kynek Mlynárce Párovce Párovské Háje Staré mesto Šúdol Zobor

Absolute frequency 1150 690 320 695 600 225 1220 75 280 180 85 80 25 45 495 75 615 45 20 40 30 240 25 105

Relative frequency (%) 62.5 37.5 17.4 37.8 32.6 12.2 66.3 4.1 15.2 9.8 4.6 4.3 1.4 2.4 26.9 4.1 33.4 2.4 1.1 2.2 1.6 13 1.4 5.7

new innovative approaches and solutions. Planning the urban transport in the long term is not easy, as there is a number of conflicting requirements, which often lead to unsystematic steps. An inclusive policy should be the solution. This trend is also presented by the European Commission that recommends the introduction of sustainable urban mobility plans that support the balanced development of all important modes of transport towards its sustainability. The city of Nitra is one of the most economically developed cities, providing the most job opportunities. Its almost 80,000 citizens make Nitra the largest city in western Slovakia outside of Bratislava. Nitra is also known as a centre of agriculture and the largest number of festivals and events takes place here. Because of its strategic geographic location, it has become one of the most important centres of business, culture, and education in the country. Thousands of tourists visit the city every year, and in combination with the students of the two universities located here, people commuting to work and the local inhabitants, transport is becoming a more topical issue [21, 22]. The city of Nitra covers an area of 100.45 km2 (38.8 sq. mi); it is currently the fifth largest city in Slovakia with a population of 78,203 (as of 30 June 2020).

112

M. Holotová et al.

3.2 European Mobility Week: A Chance to Change the Patterns of Travel Behaviour European Mobility Week (EMW) takes place annually from 16 to 22 September, with the goal of encouraging the European municipalities to enhance, promote and implement sustainable transport among local population. The EMW campaign offers an ideal chance to introduce sustainable mobility alternatives to local inhabitants and to clarify the issues that cities are facing. By participating in this activity, cities have the opportunity to bring the benefits of green transport closer to people as well as the chance to contribute to greater progress towards better mobility in Europe by introducing public multimodal transport. Local authorities are highly encouraged to use the week to test new transport measures and to get some feedback from the public. It is also a unique chance for local stakeholders to get together and discuss the various dimensions of sustainable mobility and to find new approaches to reduce transport emissions, and to plan and introduce new measures. Each EMW edition concentrates on a specific topic connected to issue of sustainable mobility. The year 2015 was a turning point, as the main idea was multimodality in transport. The main goal of this initiative was to spur people to start thinking about the range of transport options available, and to choose the right combination of mode when travelling, with the slogan “Choose. Change. Combine” [23, 24]. In connection with our marketing research, we asked the respondents if they knew and took the advantage of EMW in the city of Nitra. The results shows that 23.1% of the respondents know this campaign and try to limit the individual car transport more, up to 54.6% of the respondents register this campaign, but they do not change their transport habits and 22.3% of the respondents do not know this campaign and hear about it for the first time.

3.3 Questionnaire Research Outcomes In the initial part of our questionnaire, we asked the respondents how often they use different modes of transport to move around the city. We noticed that the transport in the city shows the elements of monomodality as the citizens do not often use the alternative modes of transport (see Fig. 2). A car was found to be the most commonly used kind of transportation in the city, as 57.1% of the respondents marked this option. Transportation networks have not attracted the attention in recent years, despite the efforts to incorporate “green” options, enabling positive lifestyle choices such as cycling and scootering commutes or better public transport networks (bus). The city does not have “sharing systems” for the alternative modes of transport, which are popular in many cities around the world. This fact is also reflected in the results in Fig. 2, as only a small percentage of the respondents indicated frequent or occasional use of bicycles, scooters or buses. Only 17.93% of the respondents use buses for public transport what is caused by poor route

Sustainable Urban Mobility–Multimodality as a Chance for Greener Cities:. . . 88.59%

90% 75% 60%

70.92%

65.49% 58.15%

57.1% 45.38% 38.86%

45% 30%

113

26.09%

23.91% 15.76% 17.93%

16.85%

15%

22.01% 12.50%

24.18% 4.89% 6.52%

4.89%

0% car

walk

bus

Often

bicycle

Occasionally

e-scooter

Taxi

Never

Fig. 2 How often do you use different modes of transport to move around the city? Source: Authors research and processing Fig. 3 Car ownership in Nitra’s households. Source: Authors research and processing

11% 9%

1 car 45%

2 cars 3 cars and more

35%

none

planning. Route networks should be designed to provide convenient links between all city points where there is demand, so that passengers can make complex journeys by using a combination of routes. The positive finding is that up to 45.38% of the citizens choose walking for personal mobility in the city often and 38.86% of the citizens choose this “healthy option” occasionally. The expansion and the dominance of individual car transport are obvious in the city of Nitra, what is also connected with the increase in the number of cars in Nitra’s households (see Fig. 3). More than 89% of households own 1 or more cars in the city of Nitra, while only 11% of households do not own a car at all. Compared to previous years, we can see an increase in the number of cars owned by Nitra’s households, and this number will grow if the city’s sustainable urban mobility policy is not resolved. The following question is related to the citizens’ opinion on the current infrastructure for the development of the alternative modes of transport in the city (Fig. 4). The results are not optimistic as the majority of the citizens (69%) clearly state that the current conditions for the development of sustainable alternative modes of transport are insufficient. Only 11% of the respondents think that the infrastructure is sufficient and 20% of them say it is difficult to judge clearly. The purpose of this

114 Fig. 4 Do you think the current infrastructure is sufficient for the development of the alternative modes of transport in the city? Source: Authors research and processing

M. Holotová et al.

20%

11% sufficient insufficient difficult to judge 69%

Fig. 5 What kind of sharing system would you like to use if it were fully functional and cost-effective? Source: Authors research and processing

43%

57%

bike sharing e-scooter sharing

question is to point out that the promotion of sustainable urban mobility and related multimodality alone is ineffective unless the suitable conditions are created for the actual implementation and functioning. The aim of the question (What kind of sharing system would you like to use if it were fully functional and cost-effective?) was to identify the respondents’ interest in the use of the alternative modes of transport based on sharing. As the example, we used the good practice of other cities where bike sharing and e-scooter sharing systems work. The question followed the idea of putting into practice the sustainable and green modes of transport that the residents would really want to use, given the distance and terrain in the city. From the results, it can be stated that both forms of transport have the potential to be used, as the preferences of the respondents are almost identical (Fig. 5). If residents under 18 were involved in the survey, we believe that the preferences for e-scooter sharing would be increased to a much greater extent. The further question provides key information for our research, as we wanted to ascertain the extent of willingness to combine different modes of transport when travelling around the city under the condition that the continuity of transport modes for any direction would be ensured. With regard to the assessment of this question, we formulated Assumption 1, where we assumed the dependence between the willingness for multimodality of transport and urban area. By determining the above-mentioned assumption, we subsequently created a contingency table (Table 2).

Sustainable Urban Mobility–Multimodality as a Chance for Greener Cities:. . . Table 2 Contingency table for urban area and willingness for multimodality

Table 3 Results of the Chi-square test of independence

Urban area Chrenová Dolné a Horné krškany Drážovce Janíkovce Klokoˇcina Kynek Mlynárce Párovce Párovské Háje Staré mesto Zobor ˇ Cermᡠn Šúdol Total

I do not know 80 0 0 30 110 10 0 5 5 40 15 20 5 320

No 40 5 5 5 80 10 5 0 10 10 5 5 5 185

115 Yes 375 20 40 40 425 25 15 35 15 190 85 55 15 1335

χ2 test Chi-square (observed value) Chi-square (critical value) DF p-value Alpha

Total 495 25 45 75 615 45 20 40 30 240 105 80 25 1840

115.169 36.415 24