Dynamic curvature path tracking control for autonomous vehicle: Experimental results

This paper discusses a new path tracking controller strategy for autonomous vehicle. Preliminary works done by researchers shows that most of the controller performs well in a straight path and normal curve. However, the problem arises in tight curve or discontinuous path. Therefore, this work aims to reduce the gap on the tight corner problem. The controller is designed based on the dynamic curvature estimation to predict the path condition and modify the wheel speed and steering wheel angle accordingly. In this technique, the curvature estimation algorithm is presented as a required input to the controller. The curvature estimation allows the vehicle to react based on the curvature profile and threshold region detection. The comparison is made based on three different path types. The experimental results on the autonomous prototype platform are discussed to show the effectiveness of the controller.