Navigating Occluded Intersections with Autonomous Vehicles Using Deep Reinforcement Learning

Abstract

Providing an efficient strategy to navigate safely through unsignaled intersections is a difficult task that requires determining the intent of other drivers. We explore the effectiveness of Deep Reinforcement Learning to handle intersection problems. Using recent advances in Deep RL, we are able to learn policies that surpass the performance of a commonly-used heuristic approach in several metrics including task completion time and goal success rate and have limited ability to generalize. We then explore a system's ability to learn active sensing behaviors to enable navigating safely in the case of occlusions. Our analysis, provides insight into the intersection handling problem, the solutions learned by the network point out several shortcomings of current rule-based methods, and the failures of our current deep reinforcement learning system point to future research directions.

A Survey of Autonomous Driving: Common Practices and Emerging Technologies

“…An interesting take on the subject is to segment image data with path proposals using a deep segmentation network [261]. Planning a safe path in occluded intersections was achieved in a simulation environment using deep reinforcement learning in [263]. The main difference between end-to-end driving and deep learning based local planners is the output: the former outputs direct vehicle control signals such as steering and pedal operation, whereas the latter generates a trajectory.…”

Safe Reinforcement Learning on Autonomous Vehicles

“…The DQN architecture is modeled after the network presented in [12]. The simulator is designed to see cars 100m in either direction.…”

“…The specific application we investigate is making a turn at an unsigned intersection. This problem was recently explored as a non-safety constrained RL domain [12] where it was noted that the learned policy, which optimized efficiency, might be disruptive to traffic vehicles in practice. The primary concerns of these maneuvers are safety and efficiency, but balancing the two is a dynamic task.…”

Safe Reinforcement Learning with Scene Decomposition for Navigating Complex Urban Environments

“…RL has been applied to autonomous braking strategies at crosswalks [5], *This work was supported by the Honda Research Institute. 1 lane changing policies [6], and intersection navigation [7], [8]. Tram et al propose a deep reinforcement learning approach with recurrent neural networks to learn how to navigate intersections with multiple vehicles with changing behaviors [7].…”