Cooperative Lane Changing via Deep Reinforcement Learning

TL;DR

This paper proposes a deep reinforcement learning approach for autonomous vehicle lane changing that emphasizes cooperative strategies to improve overall traffic flow rather than individual vehicle efficiency.

Contribution

It introduces a reward function focused on traffic efficiency and demonstrates that cooperative lane changing enhances traffic harmony and efficiency.

Findings

Cooperative strategies outperform individual-focused approaches.

Traffic efficiency improves with cooperative lane changing.

Deep RL effectively learns cooperative lane changing policies.

Abstract

In this paper, we study how to learn an appropriate lane changing strategy for autonomous vehicles by using deep reinforcement learning. We show that the reward of the system should consider the overall traffic efficiency instead of the travel efficiency of an individual vehicle. In summary, cooperation leads to a more harmonic and efficient traffic system rather than competition