Assignments for Berkeley CS285: Deep Reinforcement Learning Course

In the project, I trained an agent for a UGV to follow a target person in front of her/his. DQN is used to train the model and investigated the capabilities and limitations of using only one model in training the follow-ahead robot.