基于多参数自适应优化的智能车轨迹跟踪

Intelligent vehicles travel under various conditions. In this study, to improve the tracking accuracy, calculation speed, and vehicle stability under different working conditions, a parameter adaptive model predictive control algorithm is proposed based on different vehicle speeds and road adhesion coefficients. Vehicle stability control is added based on a linear time-varying MPC, and two control strategies are developed based on the road adhesion coefficient. On a high-adhesion-coefficient road, the prediction and control horizons arc optimized for different vehicle speeds. On a road with a low adhesion coefficient, stability control is implemented, and weight parameters are optimized based on the improved particle swarm optimization algorithm. On the premise of ensuring the tracking accuracy of the algorithm and vehicle stability,these two strategies increase computational speed. Further, a road-friction recognition algorithm based on a feedforward neural network is designed to determine the road-surface adhesion coefficient of the parameter-adaptive trajectory tracking algorithm. CarSim-Simulink is used for the co-simulation. The results reveal that the average absolute percentage error of the road recognition algorithm is 12. 77%, which is sufficient to satisfy the requirements of the multiparameter adaptive trajectory tracking algorithm. Compared with the traditional linear time-varying MPC tracking algorithm, on roads with high and low road adhesion coefficients, the transverse mean absolute error of the multiparameter adaptive trajectory tracking algorithm is reduced by 20. 7% and 24. 6% at low speeds, whereas it is decreases by 66. 2% and 50. 7% at high speeds, respectively. The computation time of the algorithm is reduced by 40. 2%. Thus, the vehicle stability is guaranteed, and the computation time is reduced. In this study, some parameters of the trajectory tracking algorithm are optimized for different vehicle speeds and road adhesion coefficients, and the adaptive prediction horizon, control horizon, and weight parameters are used to cooperatively optimize the control, thus providing a new idea for the study of trajectory tracking control under complex working conditions.