Development of an Uncertainty Estimation-Based Adaptive Path Tracking Control Algorithm for Autonomous Vehicles with Weighted Cost Function

This study proposes an adaptive path tracking control algorithm for autonomous vehicles with weighted cost function and uncertainty estimation. To design of reasonable steering control algorithm for path tracking, consideration of vehicle velocity and proper ratio of path errors in cost function is needed. The weighted cost function has been designed using weighting factor and uncertainties of the error dynamics have been estimated using Recursive Least Squares(RLS) with forgetting. Based on the weighting factor and estimated uncertainties, steering control input for path tracking has been derived with discrete injection term based on the Lyapunov direct method and the stability of the controller was analyzed. The control algorithm was constructed in Matlab/Simulink environment and performance evaluation was conducted by co-simulation of the CarMaker and Matlab/Simulink under a curved path tracking scenario. The evaluation results showed that the adaptive path tracking controller proposed in this study can make autonomous vehicle track the desired path reasonably under various velocity conditions.