Providing safety and performance guarantees for motion planning and control algorithms is a well-studied problem for robotic systems with well-known dynamics that operate in structured environments. However, when robots operate in a real-world setting where the environment is dynamic and unstructured, common assumptions used to develop the planning algorithms are no longer valid and consequently, the safety guarantees no longer hold. In this talk, I will describe my work on developing new theoretical tools for risk-aware, stochastic motion planning to account for diverse and varying uncertainty descriptions while retaining the tractability of the state-of-the-art approaches. The theory has practical and scalable implementation and has been deployed not just in controlled laboratory settings but also in complex, real-world environments.

I will provide techniques to account for uncertainty in static, extreme terrain and in dynamic environments. I will introduce a theoretical framework for motion planning while accounting for risk in a model predictive control setting. The risk-aware control policies are distributionally-robust to the uncertainty in the environment and have probabilistic guarantees for task completion and recursive feasibility. These techniques will be described in the context of my work deployed on ground robots for search and rescue operations in the DARPA Subterranean Challenge and for aerial vehicles in dynamic environments. The goal of my research is to develop verifiable algorithms for safety-critical autonomous systems that are able to perceive their dynamic and uncertain environments to enable safe and intelligent decision-making in hazardous or sensitive environments.