The paper “Markovian Dynamics Enforcer (MaDE): Feasibility Preserving Correction on Learned Dynamics Manifolds” by Kevin Yu, Tao Guo, Constantinos Antoniou and Panagiotis Angeloudis has been accepted at NeurIPS 2026. An interactive demo is available.
How many electric trucks does a mine need, and how do their costs and emissions compare with diesel? A study by Kenta Matsui, Jose Escribano and Panagiotis Angeloudis examines these questions, accounting for the charging constraints that affect fleet size.
Their paper, “Economic and environmental right-sizing of battery-electric haulage fleets in open-pit mines”, appears in the October issue of the Journal of Cleaner Production. An interactive demo is available.
Jose Escribano has taken over as programme director for the 2026–27 academic year, succeeding Panagiotis Angeloudis, who held the role for 14 years. Jose previously served as the programme’s Director of Admissions.
Josh Kirk has been awarded his PhD for his thesis, “Enhancing Resilience in Global Commodity Networks: Advanced Quantitative Methods for Supply Chain Analysis Under Geopolitical and Climate Disruptions”.
Josh joined the lab in 2023 as a Skempton Scholar. He now works at East Alpha.
Kevin Yu will be presenting “Markovian Dynamics Enforcer (MaDE): Feasibility Preserving Correction on Learned Dynamics Manifolds” at the 40th Annual Conference on Neural Information Processing Systems. MaDE corrects the output of any trajectory predictor, one step at a time, so that it stays consistent with a learned model of the vehicle’s dynamics, without retraining the predictor.
The paper was co-authored with Tao Guo, Constantinos Antoniou, and Panagiotis Angeloudis.
Panagiotis Angeloudis delivered “The Price of Efficiency: Supply Chain Risk in an Interconnected World” for Imperial Executive Education’s programme with Corndel. Managers from across industry attended the online session.
The masterclass examined how behavioural, geopolitical and structural risks spread through supply chains. It covered redundancy, diversification and hedging, alongside the use of AI and agent-based models to stress-test supply chains.
Tom He presented “A Hyper-Metaheuristic for Vehicle Routing Problem with Drones” at the 29th IEEE International Conference on Intelligent Transportation Systems (ITSC 2026), held in Naples, Italy, from 15 to 18 September.
The paper was co-authored with Jose Escribano and Panagiotis Angeloudis.
Panagiotis Angeloudis led the Maritime Futures session and moderated its panel at the first Future Vehicle Technology Symposium. The programme covered quantum positioning, nuclear propulsion and wind-assisted propulsion.
Imperial’s Vehicle Futures Hub organised the event in the City and Guilds Building, South Kensington.
Kenta Matsui presented “ADMM-based Distributed Coordination of CAVs at Unsignalised Intersections” at the 2026 European Control Conference (ECC), held in Reykjavík, Iceland, from 7 to 10 July.
Co-authored with Jose Escribano and Panagiotis Angeloudis, the paper coordinates automated vehicles at intersections shared with human drivers. Vehicles plan locally while a coordinator aligns arrival times at potential conflict points.
Keyang Zhang has passed his PhD viva for research on the game-theoretic analysis of future mobility deployments in urban logistics.
He was co-supervised by Panagiotis Angeloudis, Jose Escribano and Dario Paccagnan from Imperial’s Department of Computing. His examiner was Luciano Greco from the University of Padova.
The transport sector advisory group is convened by the Institute for the Future of Work and held its first session on 10 June.
Panagiotis Angeloudis contributes research expertise on the effects of autonomy at scale across vehicle fleets.
A paper in Transportation Research Part C: Emerging Technologies examines how to reconstruct vehicle paths through an intersection using only short observations at entry and exit. These paths support analysis of traffic interactions, delays and queues.
“Reconstructing vehicle trajectories at intersections from extremely sparse entry-exit observations” was written by Tao Guo, Kevin Yu, Panagiotis Angeloudis and Constantinos Antoniou.
Panagiotis Angeloudis delivered the session on supply chain risk management for Imperial Executive Education’s Risk Management programme, a five-day course held in person at South Kensington. The session covered supply chain vulnerabilities, disruption and resilience.
For future instances of the programme, see the Imperial Executive Education website.
Panagiotis Angeloudis gave the keynote “The Pulse of Maritime AI: Recent Developments, Sector Implications, and Ongoing Research” at SMRC x MTEC/ICMASS 2026, held during Singapore Maritime Week on 22 and 23 April.
The talk examined how domain-specific engineering and modelling can make AI useful in maritime operations, covering fleet decarbonisation, mixed autonomous and conventional traffic, and port-call sequencing.
How should road space be shared between parking and moving traffic? A study in Expert Systems with Applications examines on-street parking lane assignments that balance these demands as autonomous vehicles enter urban transport systems.
“Intelligent urban on-street parking space management for autonomous vehicles” was written by Qiming Ye, Prateek Bansal, Felix Feng, Simon Hu and Panagiotis Angeloudis.
Panagiotis Angeloudis has joined the Vehicle Futures Hub, led by Professor Gregory Offer. The Hub brings together research across Imperial’s Faculty of Engineering on sustainable, safe and reliable vehicle technologies.
Panagiotis will co-lead the Maritime Technologies research theme with Professor Michael Bluck and contribute to the Autonomous Vehicles research theme.
“Large (Vision) Language Models for Autonomous Vehicles: Current Trends and Future Directions” reviews the use of language and vision-language models in autonomous driving. It appears in IEEE Transactions on Intelligent Transportation Systems.
The authors are Hanlin Tian, Kethan Reddy, Felix Feng, Mohammed Quddus, Yiannis Demiris and Panagiotis Angeloudis.
A study in Transportation Research Part C: Emerging Technologies combines a decision tree with maritime domain knowledge to help identify higher-risk ships for inspection. The model was evaluated using inspection records from the Port of Hong Kong.
“Prediction of ship risk by a monotonic decision tree” was written by Ran Yan, Shuo Jiang, Panagiotis Angeloudis, Xinhu Cao, Jing Wang and Shuaian Wang.
Kate Jack will research resilience in future mobility systems, developing a quantitative framework to assess transport infrastructure’s readiness for combined physical and cyber risks.
Kate brings more than 20 years of leadership experience in telecommunications, energy and smart mobility. As Global Sector Lead for Smart Mobility at Stantec, she established its smart mobility business in North America.
Chenlu Kou applies computer vision and behavioural modelling to transport safety and operational efficiency. She is jointly supervised by Panagiotis Angeloudis and Yiannis Demiris.
Chenlu holds an MSc in Transport with Distinction from Imperial College London and a BEng in Traffic Engineering from Southeast University.
DeepSafe has published a report on UK public perceptions of autonomous vehicles and advanced driver assistance systems.
The report presents findings from a year-long study involving over 1,000 participants. Download the report (PDF).
Panagiotis Angeloudis spoke on maritime applications of artificial intelligence at the Singapore Maritime Research Conference 2025.
The talk drew on the lab’s research and teaching in autonomy, freight transport and artificial intelligence.
The lecture, titled “Mitigating Risk in Autonomous Vehicle Operations: Urban Autonomy and Beyond”, covered TSL’s work on AV training models based on real-world traffic incidents and camera data.
Our researchers presented work on multi-agent reinforcement learning for autonomous vehicle control at UK-MAS 2025.
Vincent Wu won the Best Presentation Award at the annual Departmental PhD Research Symposium for the second consecutive year.
Visitors used our driving simulator to explore driving in a city with autonomous vehicles and discussed the lab’s research with our team.
Felix Feng has been promoted to Research Fellow and Laboratory Manager at TSL. He will oversee the laboratory’s research infrastructure while continuing his work on autonomous vehicle safety and validation.
His research uses configurable automotive system models to simulate vehicle motion and interactions with other road users, supporting the assessment of autonomous vehicle behaviour.
TSL alumna Ellin Ivarsson has received Imperial College London’s Emerging Alumni Leader Award 2025.
Ellin has led infrastructure projects in the United Kingdom, Australia and Latin America. At the World Bank, she works on transport projects in some of Colombia’s poorest regions.
Leandro Parada has passed his PhD viva for research on multi-agent reinforcement learning for safe and efficient autonomous vehicle control. His work includes methods for detecting and isolating malicious agents that disrupt communication and coordination between vehicles.
Leandro will continue his research at TSL as part of the DeepSafe project.
Panagiotis Angeloudis delivered a talk at the 5th Athens Triennial Meeting, celebrating the 40th anniversary of the Costas Grammenos Centre for Shipping, Trade and Finance.
Discussions covered applications of artificial intelligence in maritime operations.
Imperial’s new MSc in Transport with Data Science, directed and designed by Panagiotis Angeloudis, welcomed its first cohort of students.
The curriculum covers the mathematical and statistical foundations of machine learning and their applications in transport engineering. Students learn to evaluate AI methods alongside established engineering approaches.
Panagiotis Angeloudis has been promoted to Professor of Transport Systems & Logistics at Imperial College London.
TSL contributes to Imperial College Business School’s executive education programme with CMA CGM Group at TANGRAM in Marseille. The programme covers transport and logistics decarbonisation for more than 1,000 CMA CGM employees.
Panagiotis Angeloudis delivers Day 2, “Maritime Decarbonisation Policy and Operational Levers”, covering regulations, technologies and operational measures to reduce shipping emissions.
Panagiotis Angeloudis has been appointed a Turing Fellow at the Alan Turing Institute, the UK’s national institute for data science and artificial intelligence.
Through the fellowship, he will collaborate with researchers across the UK on applications of data science and artificial intelligence in transport and logistics.
TSL alumnus Eduardo Candela has raised seed funding for MAIHEM.ai, which joined Y Combinator’s W24 cohort. The startup develops AI agents that simulate user interactions to test large language model applications during development and deployment.
Eduardo completed his PhD in AI Safety at Imperial College London in October 2023.