Autonomous Vehicle Collision Avoidance With Racing Parameterized Deep Reinforcement Learning

TL;DR

This paper develops a parameterized deep reinforcement learning collision avoidance policy for autonomous vehicles, outperforming traditional methods in simulation with significantly reduced computational requirements.

Contribution

Introduces a novel DRL-based collision avoidance policy that leverages racing overtaking without explicit trajectory guidance, achieving superior performance and efficiency.

Findings

Both DRL policies outperform MPC-APF baseline in simulation.

Reversed heading policy reduces head-to-head collision risk by 30%.

Policies transfer zero-shot to scaled hardware with 31x fewer FLOPS.

Abstract

Road traffic accidents are a leading cause of fatalities worldwide. In the US, human error causes 94% of crashes, resulting in excess of 7,000 pedestrian fatalities and $500 billion in costs annually. Autonomous Vehicles (AVs) with emergency collision avoidance systems that operate at the limits of vehicle dynamics at a high frequency, a dual constraint of nonlinear kinodynamic accuracy and computational efficiency, further enhance safety benefits during adverse weather and cybersecurity breaches, and to evade dangerous human driving when AVs and human drivers share roads. This paper parameterizes a Deep Reinforcement Learning (DRL) collision avoidance policy Out-Of-Distribution (OOD) utilizing race car overtaking, without explicit geometric mimicry reference trajectory guidance, in simulation, with a physics-informed, simulator exploit-aware reward to encode nonlinear vehicle kinodynamics. Two policies are evaluated, a default uni-direction and a reversed heading variant that navigates in the opposite direction to other cars, which both consistently outperform a Model Predictive Control and Artificial Potential Function (MPC-APF) baseline, with zero-shot transfer to proportionally scaled hardware, across three intersection collision scenarios, at 31x fewer Floating Point Operations (FLOPS) and 64x lower inference latency. The reversed heading policy outperforms the default racing overtaking policy in head-to-head collisions by 30% and the baseline by 50%, and matches the former in side collisions, where both DRL policies evade 10% greater than numerical optimal control.