Majid Khadiv

“This paper solves a critical deployment gap for legged robots — RL locomotion policies trained without safety constraints will eventually step on or collide with real-world objects, and this paper proposes a practical, modular filter that intercepts and corrects unsafe footstep plans in real-time without retraining the underlying policy.”

“Our safety filter is agnostic to the internals of π and requires only that the policy accepts contact locations as input.”

“The contact-conditioned policy framework referenced throughout as [25] is 'Omar and Khadiv (2026), Learning to act through contact: a unified view of multi-task robot learning.'”