An Independent Study of Reinforcement Learning and Autonomous Driving

TL;DR

This paper explores reinforcement learning techniques, including Q-learning and deep Q-networks, applied to autonomous driving, highlighting their effectiveness and the impact of environment configurations on training performance.

Contribution

It provides an independent analysis of reinforcement learning algorithms and their application to autonomous driving with safety constraints, including experimental insights.

Findings

Q-learning successfully applied to Taxi environment

Deep Q-network trained on Cart-Pole environment

Reinforcement learning performance affected by environment configurations

Abstract

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.