Balance Mechanism Design for the Fusion of Pure Pursuit and PI Tracking Controller

Abstract

Making use of the PID controller to make up the defeat of the geometric controller like Pure Pursuit is a very effective way to improve the effects of the Pure Pursuit (PP). However, existed methods didn't explain the internal principle of the fusion. This paper will try to give a detailed analysis about the fusion mechanism. Firstly, we will discuss the error caused by PP, and then we will separately analyze how the PI controller correct the error. Based on the above discussion, we conclude that the weight parameters are very important to the performance of the fusion controller. Thereby, we proposed a Genetic Algorithm (GA) based parameter tuning method to improve the accuracy of the fusion method. Finally, the experiments on CARLA simulator prove the effects of GA based PP_PI (PP fusion with PI) method in urban driving scenario.