Optimization of Pure Pursuit Controller based on PID Controller and Low-pass Filter

The geometric controller is widely used to solve the path tracking problem in the autonomous vehicle. However, its performance is dependent on the look-ahead distance that is difficult to be decided in the real driving. Besides, the geometric relationship isn't enough to express the impacts of the tracking error in some extreme driving situations. In this paper, we use the geometric controller Pure Pursuit as the basic controller, and add PI (Proportional Integral) controller to improve the above geometric controller's problems. We name our controller as PP_PI. Moreover, to reduce the sudden change in the steering angle and improve the steering stability, we design a low-pass filter to smooth the final output steering angle. To assess the performance of PP_PI, we compare PP_PI with Pure Pursuit and PID on the simulator. The experiment results prove PP_PI could effectively reduce the dependence on the look-ahead distance and obtain better tracking performance.