Lasse Peters
lasse-peters.net
I am a postdoctoral researcher at the ICON Lab at UC Berkeley. My research revolves around data-efficient learning for multi-agent interaction. I ground my work in real-world applications, such as multi-agent manipulation and mobile robotics.
Prior to coming to Berkeley, I completed my PhD in the department of Cognitive Robotics (CoR) at Delft University of Technology working with Prof. Javier Alonso-Mora and Prof. Laura Ferranti. My dissertation, Game-Theoretic Motion Planning for Multi-Agent Interaction, developed game-theoretic methods for planning and inference in multi-agent systems. Before that, I worked as a full-time research scholar at the Photogrammetry & Robotics Lab at University of Bonn under the supervision of Prof. Cyrill Stachniss. There, I developed methods for solving inverse games in continuous state-action spaces.
I hold a M.Sc. in Mechatronics and a B.Sc. in Mechanical Engineering from TU Hamburg, Germany, in 2020 and 2017. During that time, I was a member of the university’s RoboCup SPL team and later a visiting student and research scholar at UC Berkeley where I worked with Prof. Claire J. Tomlin and Prof. Zachary N. Sunberg at the Hybrid Systems Laboratory.
For coordinating meetings, this calendar shows times when I will be busy.
Email: lasse.peters@mailbox.org
Matrix: @lassepe:matrix.org
GitHub: github.com/lassepe
PGP: 8A1F6554BF8772E9
May, 2026. New preprint available on arXiv: “Coordinated Diffusion: Generating Multi-Agent Behavior Without Multi-Agent Demonstrations”.
May, 2026. Our paper “Controllability in Preference-Conditioned Multi-Objective Reinforcement Learning” has been accepted to NeuS 2026.
Apr, 2026. Our paper “Homotopy-Guided Potential Games for Congestion-Aware Navigation” has been accepted for publication in the IEEE Robotics and Automation Letters (RA-L).