智能车辆横向轨迹跟踪的显式模型预测控制方法

Explicit Model Predictive Control for Intelligent Vehicle Lateral Trajectory Tracking

为保证智能车辆跟踪控制的精确性、行驶稳定性和在线实时性,设计了一种显式模型预测跟踪控制方法。建立了预测时域内的跟踪精准性和行驶稳定性指标函数及约束,将跟踪控制问题转化为动态干扰下的主动转向最优化问题;为提高实时性,将传统模型预测控制系统转化为与之等价的显式多面体分段仿射系统,运用参数分区上的显式控制律求得最优转角控制量。Carsim和Simulink联合仿真结果表明,该控制方法下,位置误差均值为0.1956 m,航向角误差均值为0.276°,同时最大横向载荷转移改善5.92%,最大轮胎利用附着系数改善9.81%,且平均单步运行速度提升53.97%。

To ensure the accuracy,driving stability and online real-time of intelligent vehicle tracking control,an explicit model predictive tracking control method is designed.The cost functions and constraints for tracking accuracy and driving stability in the prediction time domain are proposed.The tracking control problem is transformed into the optimization of the active steering angle with dynamic disturbances.To improve the real-time performance,the traditional model predictive control system is transformed into an equivalent explicit polyhedral piece-wise affine(PPWA)system,and the active steering angle of front wheel is gained by the explicit law on parameter partition.Carsim and Simulink simulation results show that the mean lateral position error is 0.1956 m and the mean heading angle error is 0.276°with this method,meanwhile the maximum lateral load-conversion is improved by a rate of 5.92%and the maximum tire utilization adhesion coefficient is improved by a rate of 9.81%,and the average single-step running speed is improved by a rate of 53.97%.