PPO research on predictive action projection for constrained continuous control, including fixed-geometry and transfer evaluations.
reinforcement-learning robotics continuous-control ppo safe-reinforcement-learning safety-filter constrained-control action-projection
-
Updated
Nov 5, 2026 - Python