Robot Obstacle Avoidance Controller Based on Deep Reinforcement Learning

Figure 5

An episode in training, with the obstacle in white, the end of the road in red, and the viewpoint following the Agent, where the car hit the obstacle, terminating this episode.