Development of lateral control system for autonomous vehicle based on adaptive pure pursuit algorithm

Abstract

This paper proposes a lateral control system based on adaptive pure pursuit algorithm. The lateral control system consists of the path tracker and primitive driver. The path tracker is improved than original pure pursuit method. The original pure pursuit method is influenced by look ahead distance that is dynamically adjusted with velocity. If look ahead distance is short, tracking performance is good. But if controller has long look ahead distance, we can see cutting corner at curved path. To reduce tracking error, we applied the PI (Proportional-Integral) control theory in lateral offset. Proportional gain is constant and integral gain is a function of the road curvature. The total desired steering angle is sum of look ahead distance desired angle and lateral offset desired angle. The primitive driver controls steering actuator by using PID (Proportional-Integral-Derivative) controller. The performance of proposed path tracking algorithm is a good at curved path, but similar to original path tracking method at low curvature path.