Prof. John A McDermid OBE FREng
University of York
Title：AI, Autonomy and Assurance: the BIG Picture
Plenary Session 1
Bio:
John McDermid became Professor of Software Engineering at the University of York in 1987 and Lllyd’sRegister Foundation Chair of Safety in 2024. His research covers a broad range of issues in systems, software and safety engineering. He became Director of the Lloyd’s Register Foundation funded Assuring Autonomy International Programme in January 2018 and its successor the Centre for Assuring Autonomy in 2024, focusing on safety of robotics and autonomous systems. He has acted as an advisor to government and industry, including FiveAI, the UK Ministry of Defence and Rolls-Royce, and he was a non-executive director of the UK Health and Safety Executive from 2019 to 2024. He has been actively involved in standards development, including work on safety and software standards for civilian and defence applications; he currently chairs a British Standards Institution advisory board on standards for connected autonomous vehicles. He is author or editor of six books and has published almost 500 papers. He became a Fellow of the Royal Academy of Engineering in 2002 and was awarded an OBE in 2010.
Abstract:
The rail industry worldwide has an excellent safety record and the introduction of autonomy and artificial intelligence (AI) to assist in train operations must not degrade safety and, ideally, should improve it. This poses challenges for the assurance of AI and autonomy. As well as needing to address safety, there is a need to consider ethical issues as use of AI can introduce or reinforce biases and give rise to unjustifiable shifts of risk and responsibility.
The talk will present approaches developed at the Centre for Assuring Autonomy (CfAA) at the University of York for reasoning about the safety or AI and autonomy, and include the CfAA’s recent work on the balanced, integrated, grounded (BIG) framework which sets safety assurance in a wider ethical framework. It will also consider the potential uses of explainable AI for assurance and in other roles in managing safety. It will conclude by identifying some research questions that need to be addressed to enable more widespread and safe adoption of AI in the rail sector.
Prof. Roderick Ian Muttram FREng
Fourth Insight Ltd
Title：Human Factors and Automation
Plenary Session 1
Bio:
Roderick Muttram is system and safety engineer with over 50 years’ experience.
He joined UK track authority Railtrack in 1994 becoming the Main Board Director responsible for Safety and Standards in 1997. Chief Executive of ‘Railway Safety’ from 2000 to 2003 which became the Rail Safety and Standards Board (RSSB).
He left Network Rail in 2003 and joined Bombardier Transportation as Vice-President for Safety and Quality at both at Headquarters and the Signalling Division and was a lead auditor for critical projects worldwide.
In 2012 he left Bombardier and established Fourth Insight Ltd, an engineering consultancy.
He is a Fellow of the Royal Academy of Engineering (FREng), a Fellow of the Institution of Engineering and Technology (FIET), a Fellow of the Institution of Railway Signal Engineers (FIRSE) and a Senior Member of the Institution of Electronic and Electrical Engineers (SMIEE). He will be President of the IRSE in 2026.
His interest in autonomous systems and AI dates to the 1980’s and in HF for much longer. He has been involved in drafting several recent standards in the area.
Abstract:
Humans are error prone, and tend to make wrong assumptions, jump to conclusions and develop habitual behaviours. They can also be deceitful and deliberately hide those behaviours where they lead to adverse consequences. Those characteristics apply not only to our relationships with other humans but also critically impact our interfaces with machines. Around 50 years ago the speaker first investigated a fatal accident where human factors were a major contributory cause. It taught him the importance of avoiding system designs that induce or reinforce such behaviours; the nature of guided transport means that there are few other sectors where those lessons are as important and since joining the railway industry over 30 years ago they have been reinforced many times. As we build more intelligent machines based on Machine Learning we need to fully understand the interactions between them and human operators and users. As advisory systems become more accurate and reliable any human oversight and confirmation is highly likely to become less vigilant, particularly if measures are not included to avoid such failings. Where AI has more direct control over safety outcomes then continuous monitoring and corrective action will be required and if ‘human like’ flaws emerge then controls similar to those we use to avoid human errors will be needed. The presentation will explore these issues and highlight areas where the presenter believes further research is needed to support progress.
Prof. Koseki Takafumi
The University of Tokyo
Title：Human-Cooperative Train Automation: Japanese Unique Evolution in Response to Social Challenges
Plenary Session 2
Abstract:
Since the 1970s, Japan has been a global pioneer in commercializing driverless railway systems. However, subsequently, many railway operators prioritized ensuring safety, operational stability, and economic efficiency to cope with our high passenger density, rather than pursuing full driverless operation (GOA4).
This situation has changed dramatically in the 2020s. A declining population, an aging society, and lifestyle changes due to the pandemic have prompted a reevaluation of railways’ sustainability. Rather than simply adopting existing GOA4 technology, Japan is beginning to forge its own path based on the philosophy of “human-system collaboration.” This philosophy is symbolized by the concept of “GOA2.5,” a level positioned between GOA2 and GOA3.
This approach is not merely a technical compromise; it is a strategic choice that aims to satisfy the following complex requirements simultaneously:
1) Adaptability to diverse regions: Flexibly meeting the public transportation demands of major metropolitan areas and regional lines.
2) Economic Rationality: Achieving sustainable business operations by maximizing compatibility with existing infrastructure.
3) Respect for the working environment: Creating a work environment in which skilled train crew can take pride in their work and utilize their expertise, even amidst an aging workforce.
This lecture will introduce the unique Japanese philosophy of automation and its specific initiatives. We will explore the insights that this approach, which balances operational stability with energy efficiency and is underpinned by a human-centric philosophy, can offer for the future challenges commonly faced by railway systems worldwide. We hope this will spark a lively discussion with the audience.
Prof. Dr. Nikola Bešinović
Technische Universität Dresden
Title：Linking Modes Leads the Change: Rail-Centric Approaches to the Multimodal Transport Future
Plenary Session 2
Bio:
Prof. Dr. Nikola Bešinović is Head of the Chair of Railway Operations at the “Friedrich List” Faculty of Transport and Traffic Sciences, TU Dresden. His research focuses on developing data-driven, optimization-based methods to make railway and public transport systems more efficient, reliable, resilient, and sustainable. Key areas of his work include railway traffic management, disruption and resilience in transport networks, timetabling under growing demand, and multimodal integration. He earned his PhD in Railway Science in 2017 from TU Delft. Before moving to Dresden, he was a researcher and lecturer at Delft. His academic training includes a BSc and an MSc from the University of Belgrade. Nikola’s work has been recognized with several international awards, including distinctions from IAROR, TRAVISIONS, INFORMS, and IEEE. He is further active in international collaborations & academic service.
Abstract:
As the complexity and urgency of global mobility challenges grow, the need for cohesive, efficient, and resilient multimodal transport systems becomes increasingly clear. This keynote explores the pivotal role of rail as the structural and strategic backbone in shaping the future of integrated mobility. Grounded in the theme of “linking modes leads the change,” the talk presents a research-driven vision where rail is not merely a participant in multimodality, but a leader in orchestrating systemic integration. It delves into the multifaceted nature of multimodal systems, addressing the complexities of air-rail connectivity, suburban rail coordination, micromobility linkages, and freight integration. The presentation highlights how cutting-edge methods, ranging from network theory to mathematical optimization, enable more robust disruption management, dynamic scheduling, and improved network reachability. Emphasis is placed on translating theory into practice through decision-support tools and real-world applications, illustrating the potential for rail to anchor a more intelligent, responsive, and sustainable mobility ecosystem. This talk challenges the research community to reframe transport integration through a rail-led lens as a pathway toward more adaptive and future-ready systems.
Prof. A G Hessami
Vega Systems
Title：Advances in AI Safety & Cybersecurity
Plenary Session 3
Bio:
Ali Hessami is currently the Director of R&D and Innovation at Vega Systems, London, UK. He has an extensive track record in systems assurance and safety, security, sustainability, knowledge assessment/management methodologies. He has a background in the design and development of advanced control systems for business and safety-critical industrial applications.
Ali represents the UK on European Committee for Electrotechnical Standardization (CENELEC) & International Electrotechnical Commission (IEC) – safety systems, hardware, software and cybersecurity standards committees. In 2017 Ali joined the IEEE Standards Association (SA), for the new landmark IEEE7000 standard focused on “Addressing Ethical Concerns in System Design” and was subsequently appointed as the Technical Editor and the Chair of IEEE7000 standard. In November 2018, he was appointed as the Vice Chair and Process Architect of the IEEE’s global Ethics Certification Programme for Autonomous & Intelligent Systems (ECPAIS) developing six suites of AI Ethics Criteria and associated assessment, certification, training and expert accreditation system under IEEE brand, CertifAIEd.
Ali is the founder and chair of AI Community of Practice (CoP), a global group of altruistic experts with diverse insights and expertise in AI technologies. The CoP conducts research and publishes outcomes under the Creative Commons License aimed at regulators, public and private enterprises, governments and the benefit of humanity.
Ali is a visiting professor at London City University and Beijing Jiaotong University school of informatics. He’s a Fellow of UK-IET and RSA and Senior Life Member of IEEE.
Abstract:
This keynote will cover the latest research outcomes on Safety of Agentic AI Systems and Cybersecurity in the Age of AI. Artificial intelligence systems can be classified along a spectrum of autonomy and generality. On one end are narrow AI systems that provide specific outputs based on bounded inputs, operating as tools to augment human intelligence. On the other end is artificial general intelligence (AGI) and artificial super intelligence (ASI)– AI systems that can match or exceed human-level performance across a wide range of cognitive tasks.
Agentic AI refers to an important intermediate category: AI systems that can autonomously pursue goals, adapt to new situations, and reason flexibly about the world, but still operate in bounded domains. The main focus of the talk is on the Autonomous Decision Making and Algorithmic Learning Systems and outcome of latest research on Safety Governance of Agentic AI Systems and Governance of AI Cybersecurity.
Prof. Clive Roberts
Durham University
Title：The Future of Railway Control and Operations: Emerging Trends and Transformative Opportunities
Plenary Session 3
Bio:
Clive is Executive Dean for the Faculty of Science at Durham University. He took up this post in March 2024 following 27 years at the University of Birmingham where he still retains the role of Visiting Professor of Railway Systems.
At Birmingham, Clive was Director of the University of Birmingham’s Centre for Railway Research and Education (BCRRE) and Head of School of Engineering. He also founded and led the national UK Railway Research and Innovation Network (UKRRIN), which is a £92M academia/industry partnership supported by Research England and 14 founding industrial partners. During his time at Birmingham, BCRRE won the Queen’s Anniversary Prize for Higher Education – the highest honour in the UK’s higher education sector.
At Durham, Clive continues to lead a portfolio of research aimed at improving the performance of engineering systems, leading projects focussing on digital twinning, digitalisation and automation. His spinout company, MoniRail, has provided in-service track geometry equipment across four continents, and is currently developing quantum positioning technology for railways monitoring and potentially signalling.
Clive is a Fellow of the Institution of Railway Signalling Engineers (FIRSE), Fellow of the Institution of Engineering and Technology (FIET) and leads the IEEE Technical Committee on Railroad Systems and Applications. In 2022 he was named Innovator of the Year at the UK Technology Supply Chain Awards.
His connection with China is longstanding. He was awarded a prestigious National Science Foundation of China International Young Scientists Fellowship in 2010, and has since led projects with CRRC, as well as, Beijing, Hefei and Guangzhou Metros. He has collaborated on numerous NSFC projects, including a current project with Professor Tao Wen at Beijing Jiaotong University.
Abstract:
The railway industry stands at a transformative inflection point, where convergent technologies have the potential to reshape operational paradigms and redefine the possibilities for control and operations.
The railway control systems market is experiencing robust growth, valued at $15 billion in 2025 with projected compound annual growth rates of 7-8.1% through 2035. European Rail Traffic Management System (ERTMS) deployments are accelerating interoperability while Communication-Based Train Control (CBTC) systems are expanding globally, with the autonomous train technology market reaching $4.56 billion in 2025.
This paper examines the emerging trends that have the potential to drive fundamental changes in railway systems. The paper considers how artificial intelligence, quantum technologies, digital twin implementations, and advanced communication networks are creating unprecedented opportunities for enhanced efficiency, safety, and sustainability.
As examples, the integration of Communications-Based Train Control (CBTC) systems with AI-powered predictive analytics has the potential to revolutionise traditional metro operations, enabling dynamic optimisation of train movements and reducing headways by up to 40%. Autonomous train operations are progressing from Grade of Automation Level 2 to Level 4, with fully unattended systems demonstrating operational viability across multiple global implementations and now moving from metros to mainline scenarios. Digital twin technology is establishing comprehensive virtual replicas of railway infrastructure, facilitating real-time monitoring and predictive maintenance strategies that reduce operational costs while extending asset lifecycles. Quantum computing applications are emerging as game-changing solutions for complex optimisation challenges, with recent breakthroughs demonstrating six times improvements in solvable problem sizes for train scheduling optimisation. Quantum sensing has now been demonstrated to provide accurate position information where GPS signals are not available, which could be used for future signalling systems. The proliferation of Internet of Things (IoT) sensors throughout railway infrastructure is enabling continuous condition monitoring, with over 50,000 kilometres of track now instrumented for real-time data collection. These technological convergences are supported by 5G network implementations that provide the ultra-low latency communication essential for advanced railway automation with higher performing communications networks ready to be implemented in the near future.
These developments collectively indicate that the future of railway operations will be characterised by intelligent, interconnected systems capable of autonomous decision-making, predictive optimization, and seamless integration across multi-modal transportation networks.
Prof. Hideo Nakamura (中村 英夫)
Nihon University
Title：REVITLISING TRAIN CONTROL SYSTEM WITH DX IN JAPAN
Plenary Session 3
Abstract:
In Japan, due to the tightening of the working population, even in urban areas, some bus transportation has been reduced one after another. In anticipation of a shortage of workers, railway operators are aiming to introduce autonomous driving that does not rely on drivers to conventional railways, and JR Kyushu has developed GOA 2.5 autonomous driving based on ATS-DK, one of the existing ATP (Automatic Train Protection), and put it into practical use in March 2024. In Japan’s autonomous driving, ATP has been equipped with an automatic train control system (ATC), but this is the first time that autonomous driving with ATS, in which drivers control brakes in response to signals, is expected to progress in its spread if autonomous driving can be carried out under the existing ATS. JR Kyushu’s ATS-DK, like ATC, was a pattern-type speed scanning type ATS that could continuously check the driving speed. However, many ATS in Japan are not patterned velocity scan ATS. In order to respond to this situation, we will evolve the low-function ATS into an advanced pattern-type speed inspection function through DX, create the foundation for autonomous driving, and announce a solution that enables autonomous operation of all line sections. In Japan, a committee aimed at the future of railways has been organized and its activities have begun. Clarify the relationship between the ideal train control system that the committee aims for and the solution proposed.
Prof. Paola Pellegrini
Université Gustave Eiffel
Title：Self-organizing Railway Traffic
Plenary Session 3
Bio:
Paola Pellegrini has been working in the field of artificial intelligence and mathematical optimization for railway transportation.
Her research field covers the development of optimization approaches to effectively exploit railway infrastructure capacity, aiming to the automation of the processes at both the tactical and the operational level. In particular, she has designed and developed a state-of-the-art algorithm for the centralized real-time railway traffic management problem named RECIFE-MILP.
Paola Pellegrini is member of the Board of the International Association of Railway Operations Research (IAROR). She is member of the editorial board of Engineering Applications of Artificial Intelligence and Special Issue Editor of IET Intelligent Transport Systems.
Abstract:
Real-time railway traffic management is a challenging task that has made the object of many studies in the literature. Most of these studies aim at supplying decision support tools to dispatchers, who are in charge of making centralized decisions on train schedules and routes when small perturbations emerge. In this talk, we will discuss the change of paradigm proposed by the SORTEDMOBILITY European project, putting forward traffic self-organization. Here, intelligent trains communicate with each other and seek a consensus on how traffic shall evolve. After presenting how we merge operations research and artificial intelligence techniques, we show through microscopic simulation that the overall system performance can be preserved while letting trains decide their own behavior. Moreover, we point out critical situations that deserve further research, in the context of both centralized and self-organizing traffic management.
