Optimized potential radius reference generator algorithm for autonomous vehicle controller development

Abstract

Trajectory tracking for autonomous vehicle is one of the field that researchers pay attention. The ultimate goal for trajectory tracking is to track the pre-defined path and follow the reference path with zero steady state error. The common modules for trajectory tracking field are reference generator, controller and plant. While most of the researchers are focusing on the controller development, less work has focused on the optimized reference generator. Optimized reference generator ensures the reference input to the controller is the optimized desired points in order to develop a good controller. Therefore, this work presents the reference generator algorithm that select the best point from the road coordinate profile before being send to the controller. The method is using the vehicle potential field and the modification from Dijkstra's algorithm to generate the path. This algorithm is useful for trajectory tracking controller development. The algorithm is verified using simulation and experiment.