The model-based algorithm for autonomous vehicle path following

This paper presents a proprietary steering algorithm for path following, whose advantage lies in its ability to be applied with vehicle dynamic models of varying complexity.The proposed algorithm was implemented using models with 3 and 10 degrees of freedom, which had been previously verified.The results were compared with those obtained using the geometric pure pursuit (PP) algorithm.Both algorithms require path approximation.In this study, path approximation was conducted using B3 functions.The presented computer simulation results indicate that the proposed steering angle selection algorithm demonstrates greater accuracy than the PP algorithm.