SKooP: Symmetric Koopman Predictions for Legged Robot Reinforcement Learning

Leverages symmetric Koopman operator predictions within a reinforcement learning framework to achieve faster training and better generalization for legged robot locomotion.