Artificial Intelligence (AI) is dominating the prominent emerging technologies , especially the ones used in autonomous vehicles. Among those emerging technologies, Machine Learning, Digital Twins, Internet of Things and Self-Healing are expected to reach their plateau of productivity in less than 10 years. The Shift2Rail RAILS project 1 has its roots in the new wave of research and applications that goes under the name of Industry 4.0. This term refers to the application of machine learning and cognitive computing to leverage an effective data exchange and processing in manufacturing technologies, services and transportation , laying the foundation of what is commonly known as the fourth industrial revolution. Several industries are impacted and, although until now the ones that have benefited the most are logistics  and manufacturing , transport systems represent one of the fields in which machine learning and other techniques is expected to have a very important impact in the near future. RAILS takes up the challenge in the rail sector supporting the definition of new research directions: the ultimate goal of RAILS is to investigate the potential of AI in the rail sector in continuity with ongoing research in railways, in particular within the Shift2Rail innovation program , and to contribute to the definition of roadmaps for future research in next generation signalling systems, operational intelligence, smart-maintenance and network management. To reach its goal RAILS aims at: a) determining the gaps between AI potential, possible future scenarios and applications with the status-quo in the rail sector, in order to recognize the required innovation shifts, b) developing methodological and experimental proof-of-concepts through feasibility studies for the adoption of AI and related techniques in safety and rail automation, predictive maintenance and defect detection, traffic planning and management, c) designing transition pathways toward the rail system scenario: identification of the new research directions to improve reliability , maintainability, safety, and performance through the adoption of AI. In pursuing these objectives, RAILS wants to take a critical approach to the opportunities offered by AI in the rail sector, addressing the need for explainable, reliable and trustworthy AI technologies, to support the development of the new "Railway 4.0". 1 https://rails-project.eu/. Hence, the main ambition of RAILS is to understand and identify which approaches within the broad area of AI can have a meaningful impact on railway systems, with possible migration and technology transfer from other transport sectors and other relevant industries, like avionics, robotics and automotive, where the application of AI has proven to be feasible and advantageous. Some preliminary results described in  show that the integration of AI solutions and techniques to railways is still in its infancy despite of the work done in maintenance and traffic planning and management, as well as in investigating safety and security issues but also reveal that there is a great potential for principles driving research and real world applications in other sectors to be transferable to railways. With respect to safety related aspects, emerging threats (e.g. the so-called adversarial attacks) and certification issues could be addressed when adopting AI in autonomous and cooperative driving (e.g. virtual coupling), based on the concepts of explainable AI (XAI) and trustworthy AI. With respect to cyber-physical threat detection, innovative approaches could be developed based on AI models like Artificial Neural Networks (ANN) and Bayesian Networks together with multi-sensor data fusion and artificial vision. Resilience and optimization techniques based on genetic algorithms and self-healing could be addressed to face failures and service disruptions, as well as to increase efficiency and line capacity. Transport management problems, such as timetabling and real-time traffic rescheduling, are notoriously difficult, and commonly referred as to NP-hard problems. Recently, machine learning has been applied to solve NP-hard scheduling problems, giving a promising direction as an alternative to heuristics.
Preface The European Dependable Computing Conference (EDCC) is an international annual forum for researchers and practitioners to present and discuss their latest research results on theory, techniques, systems, and tools for the design, validation, operation, and evaluation of dependable and secure computing systems. Traditionally one-day workshops precede the main conference: the workshops complement the main conference by addressing dependability or security issues on specific application domains or by focusing in specialized topics, such as system resilience. The 18th edition of EDCC was held in Zaragoza, Spain, during September 12–15, 2022. After two years of virtual events, due to the COVID-19 pandemic, EDCC 2022 returned as a physical event, allowing researchers and practitioners to meet and exchange ideas face to face. The workshops day was held on Monday, September 12, 2022. Four workshop proposals were submitted to this 18th edition and, after a thoughtful review process led by the workshop chair, all of them were accepted. The evaluation criteria for the workshops selection included the relevance to EDCC, the timeliness and expected interest in the proposed topics, the organizers’ ability to lead a successful workshop, and their balance and synergy. These joint proceedings include the accepted papers from three of these workshops: – 14th International Workshop on Software Engineering for Resilient Systems (SERENE 2022); – 3rd International Workshop on Dynamic Risk managEment for Autonomous Systems (DREAMS 2022); – 3rd International Workshop on Artificial Intelligence for RAILwayS (AI4RAILS 2022). All these three workshops together received a total of 22 submissions. Each workshop had an independent Program Committee, which was in charge of reviewing and selecting the papers submitted to the workshop. DREAMS 2022 and AI4RAILS 2022 adopted a single-blind review process while SERENE 2022 adopted a double-blind one. All the workshop papers received three reviews per paper (67 reviews in total). Out of the 22 submissions, 11 papers were selected to be presented at the workshops (acceptance rate of 50%) and all of these 11 papers were included in these proceedings. Many people contributed to the success of the EDCC workshops day of this 18th edition. I would like to express my gratitude to all those supported this event. First, I thank all the workshops organizers for their dedication and commitment, the authors who contributed to this volume, the reviewers for their help in the paper assessment, and the workshops participants. I would also like to thank all the members of the EDCC Steering and Organizing Committees, in particular Simona Bernardi and José Merseguer (the General Chairs), who worked hard to bring back EDCC as a physical event. A special thank you to Simin Nadjm-Tehrani (the Program Chair) for her precious suggestions and Diego Peréz Palacín and Francesco Flammini (the Publicity Chairs) for disseminating the calls for papers. Finally, many thanks to the staff of Springer who provided professional support through all the phases that led to this volume.
In the last years, there has been a growing interest in the emerging concept of Digital Twins (DTs) among software engineers and researchers. DTs represent a promising paradigm to enhance the predictability, safety, and reliability of cyber-physical systems. They can play a key role in different domains, as it is also witnessed by several ongoing standardisation activities. However, several challenging issues have to be faced in order to effectively adopt DTs, in particular when dealing with critical systems. This work provides a review of the scientific literature on DTs in the railway sector, with a special focus on their relationship with Artificial Intelligence. Challenges and opportunities for the usage of DTs in railways have been identified, with interoperability being the most discussed challenge. One difficulty is to transmit operational data in real-time from edge systems to the cloud in order to achieve timely decision making. We also provide some guidelines to support the design of DTs with a focus on machine learning for railway maintenance. Keywords: Digital Twins, Railways, Artificial Intelligence, Machine Learning, Cyber-physical Systems, Internet of Things
There is growing interest in the concept of digital twins (DTs) among software engineers and researchers. As an emerging topic, DTs are a promising paradigm to enhance the predictability, safety, and reliability of cyber-physical systems in different domains. Despite the increasing trend, it is continually challenging to decide the best approach to implement DTs. Moreover, to the best of the author's knowledge, it was found that there is a lack of conducted research and no systematic reviews on DTs in the transportation industry, especially in the area of railway systems. Therefore, following the systematic literature review method, this thesis work has identified 363 articles in four digital libraries, of which 60 primary articles were included to address three research questions. The review shows that most of the reviewed articles focus on the railway subarea maintenance and inspection, the DT enabling technology artificial intelligence is the most coupled technology. An in-depth analysis found that most of the articles apply machine learning algorithms and techniques in DTs to detect faults, predict failures, make automated decisions, and monitor health status to optimize railway systems. It was also found that in-teroperability is the most discussed challenge, where the difficulty is to transmit operational data in real-time and also achieve real-time decision making. Furthermore, the analysis shows several opportunities and advantages of DTs, such as reducing maintenance costs and the positive contribution to a reduction in freight transport by road. Finally, based on the findings of the conducted review, a guideline to support the design of a DT for predictive maintenance in railways in the form of a flowchart is presented and explained.
Within Artificial Intelligence, Deep Learning (DL) represents a paradigm that has been showing unprecedented performance in image and audio processing by supporting or even replacing humans in defect and anomaly detection. The Railway sector is expected to benefit from DL applications, especially in predictive maintenance applications, where smart audio and video sensors can be leveraged yet kept distinct from safety-critical functions. Such separation is crucial, as it allows for improving system dependability with no impact on its safety certification. This is further supported by the development of DL in other transportation domains, such as automotive and avionics, opening for knowledge transfer opportunities and highlighting the potential of such a paradigm in railways. In order to summarize the recent state-of-the-art while inquiring about future opportunities, this paper reviews DL approaches for the analysis of data generated by acoustic and visual sensors in railway maintenance applications that have been published until August 31st, 2021. In this paper, the current state of the research is investigated and evaluated using a structured and systematic method, in order to highlight promising approaches and successful applications, as well as to identify available datasets, current limitations, open issues, challenges, and recommendations about future research directions.
In this paper, we present some of the results achieved during the ongoing RAILS (Roadmaps for Artificial Intelligence Integration in the Rail Sector) research project, funded by the European Union Shift2Rail Joint Undertaking, and specifically in Work Package 2, addressing AI for autonomous and cooperative driving in future smart railways. We will provide a brief state-of-the-art and opportunities for future research, focusing on the trustworthiness of intelligent train control, also based on the analysis of related transportation sectors. As for other safety-critical sectors, the use of AI for autonomous driving in railways represents a challenge. This happens when an appropriate "safety envelope" cannot be guaranteed, such as when the Automatic Train Protection (ATP) is missing or not working properly (e.g., in limited supervision operating modes), or in cases when it is specifically used to improve safety (e.g., when employed for on-track obstacle detection). In many situations, however, AI can be advantageously adopted to optimize several parameters when supporting Automatic Train Operation (ATO), while the ATP supervises safety. Furthermore, in limited or no supervision by the ATP, AI can be effectively used to support train drivers in respecting signals and keeping safe headways, i.e., as an advanced driving assistance system, in analogy with the latest developments in the automotive sector.
The progressive adoption of artificial intelligence and advanced communication technologies within railway control and automation has brought up a huge potential in terms of optimisation, learning and adaptation, due to the so-called “self-x” capabilities; however, it has also raised several dependability concerns due to the lack of measurable trust that is needed for certification purposes. This presentation provides a vision of future train control that builds upon existing automatic train operation, protection, and supervision paradigms. We defined some basic concepts for autonomous driving in digital railways, and summarised its feasibility in terms of challenges and opportunities, including explainability, autonomic computing, and digital twins. In particular, automatic train protection can act as a safety envelope for intelligent operation to optimise energy, comfort, and capacity, while intelligent protection based on signal recognition and obstacle detection can improve safety through advanced driving assistance.
The aim of this entry is to review existing publicly available and open artificial intelligence (AI) oriented datasets in different domains and subdomains of the railway sector. The contribution of this paper is an overview of AI-oriented railway data published under Creative Commons (CC) or any other copyright type that entails public availability and freedom of use. These data are of great value for open research and publications related to the application of AI in the railway sector. https://encyclopedia.pub/entry/14866
The progressive adoption of artificial intelligence and advanced communication technologies within railway control and automation has brought up a huge potential in terms of optimisation, learning and adaptation, due to the so-called “self-x” capabilities; however, it has also raised several dependability concerns due to the lack of measurable trust that is needed for certification purposes. In this paper, we provide a vision of future train control that builds upon existing automatic train operation, protection, and supervision paradigms. We will define the basic concepts for autonomous driving in digital railways, and summarise its feasibility in terms of challenges and opportunities, including explainability, autonomic computing, and digital twins. Due to the clear architectural distinction, automatic train protection can act as a safety envelope for intelligent operation to optimise energy, comfort, and capacity, while intelligent protection based on signal recognition and obstacle detection can improve safety through advanced driving assistance.
Nowadays it is widely accepted that Artificial Intelligence (AI) is significantly influencing a large number of domains, including railways. In this paper, we present a systematic literature review of the current state-of-the-art of AI in railway transport. In particular, we analysed and discussed papers from a holistic railway perspective, covering sub-domains such as maintenance and inspection, planning and management, safety and security, autonomous driving and control, revenue management, transport policy, and passenger mobility. This review makes an initial step towards shaping the role of AI in future railways and provides a summary of the current focuses of AI research connected to rail transport. We reviewed about 139 scientific papers covering the period from 2010 to December 2020. We found that the major research efforts have been put in AI for rail maintenance and inspection, while very limited or no research has been found on AI for rail transport policy and revenue management. The remaining sub-domains received mild to moderate attention. AI applications are promising and tend to act as a game-changer in tackling multiple railway challenges. However, at the moment, AI research in railways is still mostly at its early stages. Future research can be expected towards developing advanced combined AI applications (e.g. with optimisation), using AI in decision making, dealing with uncertainty and tackling newly rising cybersecurity challenges. Full text available at: https://eprints.whiterose.ac.uk/185584/
This Thesis project aims for a holistic overview of Artificial Intelligence (AI) applications in the railway industry. Our research covers diverse subdomains of railway systems such as traffic planning and scheduling, logistics and optimization, maintenance, safety and security, passenger experience, communication, and autonomous trains. The first part of this work presents a taxonomy of different terms related to AI and the railway industry. Then, we have analyzed the state of the art of AI applied to the railway industry by conducting an extensive literature review, summarizing different tasks and problems belonging to specific domains and subdomains of the railway industry and common AI-based models implemented for their solution. The existing literature reviews typically cover a limited scope either regarding specific railway subdomains or some certain aspects of AI. Within this study we present an integrated overview with special emphasis on the data used to create AI models. To achieve this, we have also conducted an extensive review on publicly available AI-oriented datasets that can be used in the different railway domains. Finally, we present a blueprint for the implementation of AI within the railway industry based on our findings. The results of our research show that the possible applications of AI in the railway sector are vast and there are many problems and tasks that can greatly benefit from it. Moreover, very different types of data are implemented to feed AI models: including not only numerical, label and image data but a wide variety of data types ranging from sound, GPS coordinates, track geometry, speed and acceleration data, data from rolling stock vibrations, knowledge from experts, log data, temperature and geological data and more. Data can also be harvested using different technologies such as IoT devices, wireless networks, smart sensors, computer-based simulations and digital twins. These and more insights are discussed in detail within this project. With this study, we want to stress that the existence of available data is one of the critical aspects of AI applications in the railway industry, and we hope to benefit researchers in the fields of computer science and the transport industry alike by providing an insight into these valuable data and information on how it can be accessed and utilized.
Blueprints for the application of AI in the Railway Industry
Preprint accepted for publication in the proceedings of "The 4th International Conference on Reliability, Safety and Security of Railway Systems" (RSSRail'22). Please cite as: Francesco Flammini, Lorenzo De Donato, Alessandro Fantechi, Valeria Vittorini. A Vision of Intelligent Train Control. Proc. 4th International Conference on Reliability, Safety and Security of Railway Systems (RSSRail’22).
Artificial Intelligence (AI) is becoming pervasive in most engineering domains, and railway transport is no exception. However, due to the plethora of different new terms and meanings associated with them, there is a risk that railway practitioners, as several other categories, will get lost in those ambiguities and fuzzy boundaries, and hence fail to catch the real opportunities and potential of machine learning, artificial vision, and big data analytics, just to name a few of the most promising approaches connected to AI. The scope of this paper is to introduce the basic concepts and possible applications of AI to railway academics and practitioners. To that aim, this paper presents a structured taxonomy to guide researchers and practitioners to understand AI techniques, research fields, disciplines, and applications, both in general terms and in close connection with railway applications such as autonomous driving, maintenance, and traffic management. The important aspects of ethics and explainability of AI in railways are also introduced. The connection between AI concepts and railway subdomains has been supported by relevant research addressing existing and planned applications in order to provide some pointers to promising directions.
The availability of large amounts of data is crucial to the adoption of Artificial Intelligence (AI) and Machine Learning in many domains, including railways, where experimentation with those approaches started recently. Within learning paradigms, the so-called Transfer Learning allows good performance with small datasets, and therefore we believe that could contribute to a faster take-up of AI in the rail sector. To support that claim, we present a Deep Learning case study based on Transfer Learning, addressing the automatic recognition of warning bell audio signals at rail level crossings.
In recent years, Artificial Intelligence (AI) is becoming increasingly present in the daily lives of individual citizens. So much so that, in many areas there is an increasing effort to invest in research in order to create and integrate "intelligent systems" within production processes, services, inspections, etc. In this thesis, we will analyse only a small part of the immense suite of tools and disciplines that Artificial Intelligence encompasses; in particular, we will focus mainly on some aspects of Deep Learning. The work described in this Thesis has been conducted at the Department of Electrical Engineering and Information Technologies of the University of Naples Federico II, in collaboration with the research group currently working at the H2020 Shif2Rail project RAILS (Roadmaps for AI integration in the raiL Sector) and a railway company; as such the thesis addresses topics related to the application of Deep Learning to railways, and specifically to maintenance and defect detection. The purpose of this work is twofold: a) to assess the status of research on the usage of Deep Learning techniques in railway applications for maintenance purposes, with a specific focus on video analysis and audio detection, and b) to investigate the application of Deep Learning to a real world monitoring and maintenance scenario. At this aim, we have laid the groundwork to build a robust audio alarm detection system at level crossings. Hence, after a brief introduction to Machine Learning and Deep Learning (Chapter 1), the Thesis presents a Systematic Literature Review on Deep Learning for video analysis and audio detection in railways (Chapter 2) and proposes a modular system to monitor and evaluate the health status of level crossings; as a first step, a deep neural network is implemented that is able to discern between audios related to the level crossing alarm and audios related to similar sounds coming from the surrounding environment (Chapter 3). The choice to leverage only on video and audio data is due to the necessity of using non-invasive sensors so as not to interfere with the behaviour of a safety-critical system in operation.
The slides were published under one of the deliverable outputs of the RAILS project, which have been presented at the 31st European Conference on Operational Research (2021, Athens, Greece). The corresponding deliverable package 1.2 gives a comprehensive survey on the state-of-the-art applications of artificial intelligence in railway sectors. Where an in-depth statistical analysis for the distribution of surveyed papers was conducted.
This report identifies current and potential application areas of AI across railway domains. With this aim the document surveys the main railway problems to which artificial intelligence techniques are currently being applied or that could benefit from AI based approaches, as they emerged from the activities carried out in WP1 (State-of-the-art of Artificial Intelligence in railway transport). The main challenges to be faced and some steps necessary to an effective take up of AI in railways are also delineated, basing on the results of the mentioned activities, which include results from a survey, the analysis of scientific literature and relevant projects, and feedback and information from the Advisory Board. The report also provides a matching between the set of relevant problems and the current AI techniques, as well as some guidelines to drive the choice of appropriate AI approaches focusing on a particular aspect of the problem under analysis.
The aim of this paper is to review existing publicly available and open artificial intelligence (AI) oriented datasets in different domains and subdomains of the railway sector. The contribution of this paper is an overview of AI-oriented railway data published under Creative Commons (CC) or any other copyright type that entails public availability and freedom of use. These data are of great value for open research and publications related to the application of AI in the railway sector. This paper includes insights on the public railway data: we distinguish different subdomains, including maintenance and inspection, traffic planning and management, safety and security and type of data including numerical, string, image and other. The datasets reviewed cover the last three decades, from January 1990 to January 2021. The study revealed that the number of open datasets is very small in comparison with the available literature related to AI applications in the railway industry. Another shortcoming is the lack of documentation and metadata on public datasets, including information related to missing data, collection schemes and other limitations. This study also presents quantitative data, such as the number of available open datasets divided by railway application, type of data and year of publication. This review also reveals that there are openly available APIs—maintained by government organizations and train operating companies (TOCs)—that can be of great use for data harvesting and can facilitate the creation of large public datasets. These data are usually well-curated real-time data that can greatly contribute to the accuracy of AI models. Furthermore, we conclude that the extension of AI applications in the railway sector merits a centralized hub for publicly available datasets and open APIs.
This report presents a comprehensive review of research projects and state-of-the-art scientific papers of Artificial Intelligence (AI) in railway industry. It represents the output of RAILS Work Package 1 Tasks 1.2 and 1.3. The focus is on European and overseas projects mainly in United States and China. Specific emphasis is devoted to reviewing projects funded by the European Shift2Rail Joint Undertaking, which represents one of the main funding bodies in EU for railway research and innovation. We address the projects and scientific papers from a holistic railway perspective covering selected areas as defined in Deliverable 1.1, including (1) maintenance and inspection, (2) safety and security, (3) autonomous driving and control, (4) transport planning and management, (5) revenue management, (6) transport policy and (7) passenger mobility. As such, this report makes an initial step towards shaping the role of AI in future railways, and provides an in-depth summary of current focus of AI research connected to rail transport. In addition, the report determines some promising research directions towards further uptake of AI in railways. The report recognizes that the major research efforts have been put in AI for rail maintenance and inspection, while very limited or no research has been done on AI for rail transport policy and revenue management. The remaining subdomains received mild to moderate attention. The potential of AI applications is evident, but it is also highlighted that AI research in railways is at its early stages. Thus, many open research topics are envisioned, some of which could contribute to fundamental usage of AI in general. Future research can be expected towards, among others, developing advanced combined AI applications, using AI in decision making, dealing with uncertainty and tackling newly rising cybersecurity challenges. This review provides general guidelines to support railway researchers to assess and understand the usability of AI techniques. It is also meant to support industry stakeholders to promptly determine promising AI domains for given railway problems. The objective of this deliverable is to further contribute to bridging the gap between AI and railway-domain experts.
Autonomous, or self-driving, cars are emerging as the solution to several problems primarily caused by humans on roads, such as accidents and traffic congestion. However, those benefits come with great challenges in the verification and validation (V&V) for safety assessment. In fact, due to the possibly unpredictable nature of Artificial Intelligence (AI), its use in autonomous cars creates concerns that need to be addressed using appropriate V&V processes that can address trustworthy AI and safe autonomy. In this study, the relevant research literature in recent years has been systematically reviewed and classified in order to investigate the state-of-the-art in the software V&V of autonomous cars. By appropriate criteria, a subset of primary studies has been selected for more in-depth analysis. The first part of the review addresses certification issues against reference standards, challenges in assessing machine learning, as well as general V&V methodologies. The second part investigates more specific approaches, including simulation environments and mutation testing, corner cases and adversarial examples, fault injection, software safety cages, techniques for cyber-physical systems, and formal methods. Relevant approaches and related tools have been discussed and compared in order to highlight open issues and opportunities.
Autonomous, or self-driving, cars are emerging as the solution to several problems primarily caused by humans on roads, such as accidents and traffic congestion. However, those benefits come with great challenges in the verification and validation (V&V) for their safety assessment. In fact, due to the possibly unpredictable nature of Artificial Intelligence (AI), its use in autonomous cars creates concerns that need to be addressed using appropriate V&V processes that are able to address trustworthy AI and thus achieve a safe autonomy. In this study, the research work in the last ten years is reviewed to investigate the state-of-the-art in the software V&V of safe autonomous cars and summarize open issues in this field. Relevant papers have been found and classified using a systematic approach. By appropriate inclusion and exclusion criteria, a subset of primary studies have been selected for more in-depth analysis. The applicability of current approaches within reference standards such as the ISO 26262 has also been reviewed. Finally, the review has investigated the adoption of formal methods with a focus on cyber-physical systems, as well as more conventional software verification approaches such as simulation and fault injection, together with mutation testing of machine learning systems, with comparison between relevant approaches.
Technology development in the field of the Internet of Things (IoT) and more specifically in Low-Power Wide-Area Networks (LPWANs) has enabled a whole set of new applications in several fields of Intelligent Transportation Systems. Among all, smart-railways represents one of the most challenging scenarios, due to its wide geographical distribution and strict energy-awareness. This paper aims to provide an overview of the state-of-the-art in LPWAN, with a focus on intelligent transportation. This study is part of the RAILS (Roadmaps for Artificial Intelligence integration in the raiL Sector) research project, funded by the European Union under the Shift2Rail Joint Undertaking. As a first step to meet its objectives, RAILS surveys the current state of development of technology enablers for smart-railways considering possible technology transfer from other sectors. To that aim, IoT and LPWAN technologies appear as very promising for cost-effective remote surveillance, monitoring and control over large geographical areas, by collecting data for several sensing applications (e.g., predictive condition-based maintenance, security early warning and situation awareness, etc.) even in situations where power supply is limited (e.g., where solar panels are employed) or absent (e.g., installation on-board freight cars).
This book constitutes refereed proceedings of the Workshops of the 16th European Dependable Computing Conference, EDCC: Workshop on Articial Intelligence for Railways, AI4RAILS 2020, Worskhop on Dynamic Risk Management for Autonomous Systems, DREAMS 2020, Workshop on Dependable Solutions for Intelligent Electricity Distribution Grids, DSOGRI 2020, Workshop on Software Engineering for Resilient Systems, SERENE 2020, held in September 2020. Due to the COVID-19 pandemic the workshops were held virtually. The 12 full papers and 4 short papers were thoroughly reviewed and selected from 35 submissions. The workshop papers complement the main conference topics by addressing dependability or security issues in specic application domains or by focussing in specialized topics, such as system resilience.
This deliverable is the output of RAILS Work Package 1 Task 1.1. As such, it provides a taxonomy of Artificial Intelligence (AI) approaches that are capable of analyzing, predicting and improving railway systems. In this document we make a distinction between different AI perspectives addressing areas such as knowledge representation, reasoning under uncertainties, planning, machine learning, computer vision, language processing, etc. The deliverable aims at providing a taxonomic overview of relevant AI concepts to support decisions about which AI techniques would be most appropriate in order to tackle the challenges associated to modern smart-railways. In addition, several advanced AI concepts such as “trustworthy AI” and AI ethics are introduced; the European directives and resolutions are considered in taking into account the ethical dimension of AI when investigating AI for railway systems. The taxonomy is the first step towards providing a general framework to support railway decision-makers to assess and understand the usability of AI-based approaches and to support industry stakeholders to promptly determine promising AI solutions to solve certain railway problems. The objective of this deliverable is to contribute in bridging the gap between AI and railway-domain experts in terms of basic concepts and terminology.
Technology development in the field of the In-ternet of Things (IoT) and more specifically in Low-Power Wide-Area Networks (LPWANs) has enabled a whole set of new applications in several fields of Intelligent Transportation Systems. Among all, smart-railways represents one of the most challenging scenarios, due to its wide geographical distribution and strict energy-awareness. The aim of this paper is to provide an overview of the state-of-the-art in LPWAN, with a focus on intelligent transportation. This study is part of the RAILS (Roadmaps for Artificial Intelligence integration in the raiL Sector) research project, funded by the European Union under the Shift2Rail Joint Undertaking. As a first step to meet its objectives, RAILS surveys the current state of development of technology enablers for smart-railways considering possible technology transfer from other sectors. To that aim, IoT and LPWAN technologies appear as very promising for cost-effective remote surveillance, monitoring and control over large geographical areas, by collecting data for several sensing applications (e.g., predictive condition-based maintenance, security early warning and situation awareness, etc.) even in situations where power supply is limited (e.g., solar panels) or absent (e.g., installation on-board freight cars).
The European RAILS research project is investigating the potential applications of artificial intelligence (AI) within the rail sector, and helping define roadmaps for future research in next generation signalling systems, operational intelligence, and network management. The overall objective of the Roadmaps for AI integration in the raiL Sector (RAILS) research project is to investigate the potential of artificial intelligence (AI) in the rail sector, and contribute to the definition of roadmaps for future research in next generation signalling systems, operational intelligence, and network management. RAILS will address the training of doctoral students to support the research capacity in AI within the rail sector across Europe by involving research institutions with a combined background in both computer science and transportation systems, in four countries: Italy, United Kingdom, the Netherlands, and Sweden. RAILS will produce knowledge, ground-breaking research and experimental proof-of-concepts for the adoption of AI in rail automation, predictive maintenance and defect detection, traffic planning, and capacity optimisation. As such, RAILS will effectively contribute to the design and implementation of smarter railways. To this end, RAILS will combine AI paradigms like machine learning with the Internet of Things (IoT) in order to leverage on the big amount of data generated by smart sensors and applications. The research activities will be conducted in continuity with ongoing research in railways, in particular within the Shift2Rail innovation program, and will be based on in-depth analysis of AI applications in transport and other relevant sectors in order to perform a transferability study of available results to railways. Link to full paper: https://ercim-news.ercim.eu/en121/r-i/roadmaps-for-ai-integration-in-the-rail-sector-rails