Autonomous Driving in Reality with Reinforcement Learning and Image Translation

TL;DR

This paper proposes a novel reinforcement learning framework combined with image semantic segmentation to bridge the gap between virtual training environments and real-world autonomous driving, enabling safer and more effective real-world deployment.

Contribution

It introduces a new framework that integrates semantic segmentation with reinforcement learning to improve transferability from simulation to reality in autonomous driving.

Findings

Agent trained in TORCS shows improved adaptability to real-world scenarios.

Semantic segmentation enhances the transferability of reinforcement learning models.

The approach reduces the risk of accidents during real-world deployment.

Abstract

Supervised learning is widely used in training autonomous driving vehicle. However, it is trained with large amount of supervised labeled data. Reinforcement learning can be trained without abundant labeled data, but we cannot train it in reality because it would involve many unpredictable accidents. Nevertheless, training an agent with good performance in virtual environment is relatively much easier. Because of the huge difference between virtual and real, how to fill the gap between virtual and real is challenging. In this paper, we proposed a novel framework of reinforcement learning with image semantic segmentation network to make the whole model adaptable to reality. The agent is trained in TORCS, a car racing simulator.