critic

New SKooP method boosts reinforcement learning for robot locomotion

Researchers have developed SKooP (Symmetric Koopman Predictions), a novel approach to enhance reinforcement learning for legged robot locomotion. This method combines morphological symmetries with a Koopman model learne…