Local Path Planning for Autonomous Vehicles: Crash Mitigation

A path planning approach to generate a path which mitigates the effects of an inevitable crash for autonomous vehicles is presented in this brief. The model predictive control algorithm is adopted here for path planning. The artificial potential field, which describes the obstacles and the potential crash severity, are added to the control objectives to avoid the obstacle, and also to mitigate the inevitable crash. The vehicle dynamic is also considered as an optimal control objective. Based on the analysis above, the model predictive controller can guarantee the command following, obstacle avoidance, vehicle dynamics, and mitigate the inevitable crash. Simulation results verified that the proposed MPC has the abilities of obstacles avoidance and mitigation of the inevitable crash.