Learning Decision-making Behaviors from Demonstrations based on Adversarial Inverse Reinforcement Learning