Imminent Collision Mitigation with Reinforcement Learning and Vision
This addresses safety for autonomous vehicles and drivers by mitigating injury in unavoidable collisions, but it is incremental as it builds on existing reinforcement learning and injury models.
This work tackles the problem of reducing collision severity in imminent on-road crashes by using reinforcement learning to control velocity and steering, showing that policies trained with injury-based penalties outperform a straight-line braking baseline.
This work examines the role of reinforcement learning in reducing the severity of on-road collisions by controlling velocity and steering in situations in which contact is imminent. We construct a model, given camera images as input, that is capable of learning and predicting the dynamics of obstacles, cars and pedestrians, and train our policy using this model. Two policies that control both braking and steering are compared against a baseline where the only action taken is (conventional) braking in a straight line. The two policies are trained using two distinct reward structures, one where any and all collisions incur a fixed penalty, and a second one where the penalty is calculated based on already established delta-v models of injury severity. The results show that both policies exceed the performance of the baseline, with the policy trained using injury models having the highest performance.