An End-to-End Collaborative Learning Approach for Connected Autonomous Vehicles in Occluded Scenarios

TL;DR

This paper introduces a collaborative learning approach for connected autonomous vehicles using V2V networks to share LiDAR features, enabling safer navigation in occluded scenarios through end-to-end training with Proximal Policy Optimization.

Contribution

It presents a novel multi-agent control method that learns directly from experience, sharing compressed perception data to improve safety and efficiency in occluded environments.

Findings

Outperforms independent reinforcement learning methods.

Effective in real-time occluded intersection scenarios.

Maintains bandwidth efficiency while sharing perception features.

Abstract

Collaborative navigation becomes essential in situations of occluded scenarios in autonomous driving where independent driving policies are likely to lead to collisions. One promising approach to address this issue is through the use of Vehicle-to-Vehicle (V2V) networks that allow for the sharing of perception information with nearby agents, preventing catastrophic accidents. In this article, we propose a collaborative control method based on a V2V network for sharing compressed LiDAR features and employing Proximal Policy Optimisation to train safe and efficient navigation policies. Unlike previous approaches that rely on expert data (behaviour cloning), our proposed approach learns the multi-agent policies directly from experience in the occluded environment, while effectively meeting bandwidth limitations. The proposed method first prepossesses LiDAR point cloud data to obtain meaningful features through a convolutional neural network and then shares them with nearby CAVs to alert for potentially dangerous situations. To evaluate the proposed method, we developed an occluded intersection gym environment based on the CARLA autonomous driving simulator, allowing real-time data sharing among agents. Our experimental results demonstrate the consistent superiority of our collaborative control method over an independent reinforcement learning method and a cooperative early fusion method.