An Independent Study of Reinforcement Learning and Autonomous Driving
This is an incremental project for students or beginners in reinforcement learning, focusing on educational applications in autonomous driving.
The study tackled applying reinforcement learning to autonomous driving by implementing Q-learning and deep Q-network algorithms on standard environments and training a basic autonomous driving agent with safety constraints, exploring the impact of reward functions on performance.
Reinforcement learning has become one of the most trending subjects in the recent decade. It has seen applications in various fields such as robot manipulations, autonomous driving, path planning, computer gaming, etc. We accomplished three tasks during the course of this project. Firstly, we studied the Q-learning algorithm for tabular environments and applied it successfully to an OpenAi Gym environment, Taxi. Secondly, we gained an understanding of and implemented the deep Q-network algorithm for Cart-Pole environment. Thirdly, we also studied the application of reinforcement learning in autonomous driving and its combination with safety check constraints (safety controllers). We trained a rough autonomous driving agent using highway-gym environment and explored the effects of various environment configurations like reward functions on the agent training performance.