End-to-End Model-Free Reinforcement Learning for Urban Driving using Implicit Affordances

TL;DR

This paper introduces a novel end-to-end model-free reinforcement learning approach using implicit affordances, enabling an RL agent to handle complex urban driving tasks including traffic light detection, and demonstrates success in a competitive benchmark.

Contribution

The paper presents the first RL agent capable of complex urban driving tasks with traffic light detection using implicit affordances, advancing RL applications in autonomous driving.

Findings

Successfully handled lane keeping, pedestrian and vehicle avoidance, and traffic light detection.

Won the Camera Only track of the CARLA challenge.

Demonstrated effectiveness of the proposed RL method in complex urban scenarios.

Abstract

Reinforcement Learning (RL) aims at learning an optimal behavior policy from its own experiments and not rule-based control methods. However, there is no RL algorithm yet capable of handling a task as difficult as urban driving. We present a novel technique, coined implicit affordances, to effectively leverage RL for urban driving thus including lane keeping, pedestrians and vehicles avoidance, and traffic light detection. To our knowledge we are the first to present a successful RL agent handling such a complex task especially regarding the traffic light detection. Furthermore, we have demonstrated the effectiveness of our method by winning the Camera Only track of the CARLA challenge.