Machine learning for robot control

This talk explores how machine learning can enable adaptive reactive intelligence in robotic systems, addressing the fundamental challenge that robots need both reactive and deliberative capabilities. I present our research across three areas:

1. Reinforcement learning for robot control with stability and safety guarantees;
2. System identification, including adaptive control, diffusion models for dynamics learning, and computation-aware methods for edge deployment;
3. Multi-agent coordination, from zero-shot cooperation to large-scale formation control with up to 10,000 agents.

Our approaches have been validated across various platforms—from quadrotors to quadrupeds and from aerial manipulators to spacecraft rendezvous—demonstrating that successful robot learning requires both theoretical rigor and practical constraints for safe and scalable deployment.

Biography

Wei Pan is an Associate Professor in Machine Learning at the Department of Computer Science and a member of the Center for AI Fundamentals and the Center for Robotics and AI, The University of Manchester, UK. He leads teaching in UKRI AI CDT on Decision Making for Complex Systems. He is an ELLIS Scholar in the ELLIS Unit in Manchester. Before that, he was an Assistant Professor in Robot Dynamics at the Department of Cognitive Robotics and co-director of Delft SELF AI Lab, TU Delft, Netherlands and a Project Leader at DJI, China. He is an Area Chair or Associate Editor of IEEE T-RO, IEEE RA-L (outstanding AE award), ACM TOPML, RSS, CoRL, L4DC, ICRA, IROS. He received his PhD degree from Imperial College London. He is also an advisor for several robotics startups.

Host or Publisher LSE Department of Statistics