My research focuses on developing new theoretical foundations for 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.
Conformal Prediction (CP) is a statistical technique for quantifying the uncertainty of data-driven models in a distribution-free manner. We utilize these techniques for ensuring a statistically guaranteed level of task success while,
Adapting to distribution shifts in motion planning with unknown dynamic agents, and
Aligning the uncertainty of LLM-based planners by allowing the robot to ask for help when the language instruction is unclear.
Synthesis and evaluation of increasingly dynamic quadrupedal locomotion gaits like walking, trotting, and hopping for navigation over unknown terrain with a focus on obstacle avoidance using the walking gait and vision.