Aerial and Ground Robot Racing

This project explores the physical limits of ground and aerial robots. When operating robots in these regimes, unknown dynamic effects (for example, unmodelled robot dynamics or unmodelled interactions with the environment) can significantly corrupt the robot’s performance. To compensate for such effects, we apply learning-based control approaches that adapt both the reference trajectory and the feedback controller over time. The result is a significantly better tracking accuracy at increased speeds.

 

Related Publications

University of Toronto Institute for Aerospace Studies