Safe Whole-body Intelligent Robotics Lab (SWIRL) pushes the boundaries of robot learning by developing safe, intelligent systems that leverage their entire physical embodiment. We operate at the exciting intersection of online reinforcement learning, offline reinforcement learning, and imitation learning, with a relentless focus on sample efficiency and safety.

Our vision is simple yet ambitious: create robots that learn efficiently, act intelligently, and operate safely in complex, dynamic environments alongside humans.