Abstract

Reinforcement Learning (RL) aims at learning an optimal behavior policy fromits own experiments and not rule-based control methods. However, there is no RLalgorithm yet capable of handling a task as difficult as urban driving. Wepresent a novel technique, coined implicit affordances, to effectively leverageRL for urban driving thus including lane keeping, pedestrians and vehiclesavoidance, and traffic light detection. To our knowledge we are the first topresent a successful RL agent handling such a complex task especially regardingthe traffic light detection. We demonstrate the effectiveness of our method bybeing one of the top teams of the camera only track of the CARLA challenge.

 

Quick Read (beta)