Inverse Reinforcement Learning with Locally Consistent Reward Functions †, Kian Hsiang Low †, and Patrick Jaillet