The Poonawala Lab works on assured learned controllers of autonomous systems, with a particular focus on end-to-end systems where raw sensor data drives control decisions directly. We develop the stability conditions, data principles, and certification methods — rooted in dynamics and control theory — that make reliability in these systems something that can be reasoned about rather than merely observed. We apply our methods to autonomous ground robots, robotic manipulation, and manufacturing process control.

More information can be found on our research page.