Physical
Active Learning
Laboratory

Stanford University · Mechanical Engineering

We build robot learning algorithms that work in real time, in the real world, on real hardware. We target diverse platforms, aiming to enrich how robots interact and adapt in the physical world.

Follow Us Email

Overview

The Physical Active Learning Laboratory is part of the Department of Mechanical Engineering and the Stanford Robotics Center. We build algorithms that allow robots to safely learn, rapidly adapt, and actively exploit their physical environment in real time. From pushing the limits of high-performance autonomous racing to navigating complex field logistics, our mission is to deploy reliable, safety-critical learning algorithms that fundamentally enrich robot-environment interaction across diverse deployments. We leverage techniques from reinforcement learning, optimal control, dynamical systems, information theory, and more to develop algorithms that are theoretically grounded and practically deployable.

News

  • 2026

    NEW Prof. Berrueta joins Stanford University as an Assistant Professor of Mechanical Engineering, launching the Physical Active Learning Lab. Positions are available, see our join page.

  • 2026

    Our paper on environment-aware GNSS covariance learning for autonomous racing is accepted to ICRA 2026, and our workshop paper on online continual learning for robust LiDAR perception at racing speeds is named a Best Paper Finalist.

  • 2025

    Prof. Berrueta is named Associate Editor of IEEE Robotics and Automation Letters (RA-L), handling Aerial & Field Robotics and Theoretical Foundations.

  • 2025

    As technical lead of Caltech Racer, the autonomous IndyCar program makes its debut at its first U.S. road-course challenge — featured by Caltech Magazine.

  • 2025

    Invited talks: “Robot learning on the edge” at the ELLIIT Robot Learning Symposium (Lund University, Sweden), and a robotics panel at a16z Tech Week (Los Angeles).

  • 2024

    Maximum Diffusion Reinforcement Learning published in Nature Machine Intelligence, and Prof. Berrueta is named a Microsoft Future Leader in Robotics and AI. Article

People

Thomas A. Berrueta

Group Leader

Prof. Thomas A. Berrueta

Assistant Professor of Mechanical Engineering

Thomas is an Assistant Professor of Mechanical Engineering at Stanford University, where he leads the PAL Lab. Before joining Stanford, he was a Postdoctoral Researcher at Caltech's Computing and Mathematical Sciences Department, and prior to that he did his Ph.D. in Mechanical Engineering at Northwestern, where he was named a Presidential Fellow.