Welcome to the PRACTICE (Probabilistic Robotics and Control Theory in Complex Environments) Lab. Our lab develops verifiable algorithms for safety-critical autonomous systems. These systems must be able to perceive their dynamic and uncertain environments to enable safe and intelligent decision-making in hazardous or sensitive environments such as in search and rescue operations, extraterrestrial exploration, and urban driving.

Towards this goal, we have developed new theoretical foundations for risk-aware, stochastic motion planning to account for diverse and varying uncertainty descriptions while retaining the tractability of state-of-the-art approaches. This research has been used for search and rescue tasks in the DARPA Subterranean Challenge as a part of the Jet Propulsion Lab Team CoSTAR’s solution. We have developed tools for uncertainty quantification of learning-based systems such as perception systems, large and vision language model-based planners to ensure safe and reliable planning.