Path Planning for Collision Avoidance Based on Artificial Potential Field with Vehicle Dimension Constraint

In this paper, a path planning algorithm combined the artificial potential field with the vehicle dimension constraint is proposed for autonomous vehicles to generate a collision-free path. To describe the vehicle dimension constraint, an ellipse inflated model of the obstacle is constructed, and the parameters of model are determined with the minimum initial longitudinal safety distance in the lane changing process. Then, the artificial potential field of the surrounding environment is established, which includes inflated obstacle vehicle, lane line, and virtual target point. Collision avoidance path points are produced in the artificial potential field by the potential gradient descent algorithm, and optimized by a fifth-order polynomial. Simulation results demonstrate that the path generated by this method is continuous, smooth, and responds rapidly to dynamic environments.