Path-Integral-Based Reinforcement Learning Algorithm for Goal-Directed Locomotion of Snake-Shaped Robot

Figure 3

Goal-directed locomotion that the robot does not need to climb obstacles from different starting positions.