On the Development of a Path Tracking Controller by combining Optimal Preview Control and Pursuit Control Methods

Abstract

Path tracking controllers are an important part when developing self-driving vehicles. When designing such a controller for a single-track vehicle, a major challenge is that the vehicle not only has to follow a predefined path but has to be stabilized at the same time. Especially if the vehicle has slow steering dynamics (e.g. τ > 1.5s) and the path contains tight corners which should not be cut (e.g. directional change of more than 90° in less than 2s), the design of a suitable path tracking controller can be difficult. In this work an optimal preview controller that stabilizes a bicycle and that can track a given yaw rate trajectory is combined with a quintic polynomial pursuit controller. With this approach, we can achieve good tracking performance for the predefined paths, which can be seen in the simulation and experimental results.