Reinforcement Learning and Deep Learning based Lateral Control for Autonomous Driving

TL;DR

This paper presents a vision-based autonomous driving system combining deep learning for perception and reinforcement learning for control, validated in a new simulation environment, VTORCS, outperforming traditional controllers.

Contribution

It introduces a modular approach separating perception and control, and proposes VTORCS, a novel simulation environment for training and testing reinforcement learning-based driving agents.

Findings

The reinforcement learning controller outperforms LQR and MPC controllers.

The perception module demonstrates promising accuracy in track feature prediction.

The system effectively controls vehicle along the track center using only visual input.

Abstract

This paper investigates the vision-based autonomous driving with deep learning and reinforcement learning methods. Different from the end-to-end learning method, our method breaks the vision-based lateral control system down into a perception module and a control module. The perception module which is based on a multi-task learning neural network first takes a driver-view image as its input and predicts the track features. The control module which is based on reinforcement learning then makes a control decision based on these features. In order to improve the data efficiency, we propose visual TORCS (VTORCS), a deep reinforcement learning environment which is based on the open racing car simulator (TORCS). By means of the provided functions, one can train an agent with the input of an image or various physical sensor measurement, or evaluate the perception algorithm on this simulator. The trained reinforcement learning controller outperforms the linear quadratic regulator (LQR) controller and model predictive control (MPC) controller on different tracks. The experiments demonstrate that the perception module shows promising performance and the controller is capable of controlling the vehicle drive well along the track center with visual input.