MPC算法与纯跟踪算法的横向控制策略研究对比

Comparative study of lateral control strategies of MPC algorithm and PP algorithm

为探究智能车辆横向控制算法的性能差异及其适用场景,分别构建了基于预瞄模型的纯跟踪算法(PurePursuit,PP)和基于运动学模型的模型预测控制算法(Model Predictive Control,MPC),利用CarSim和Simulink对双移线和低速、中速以及高速行驶工况进行联合仿真,对比不同工况下的横向控制效果。结果表明:PP算法在低中速时具有较高跟踪精度,在高速时则有很大的误差;MPC算法在低、中、高速工况下的跟踪效果均良好,但实时性较弱。考虑到计算量和实时性,在低速工况下优先选用PP算法,中高速工况下选用MPC算法最佳。

In order to investigate the performance differences of the lateral control algorithms of intelligent vehicles and their applicable scenarios,Pure Pursuit algorithms(PP)based on the pre-scanning model and Model Predictive Control algorithm(MPC)based on the kinematic model were constructed,respectively,and CarSim and Simulink simulated jointly and compared the effects of lateral control under the working conditions of double-shift lines and low,medium,and high speeds,and the lateral control effects were compared.The results showed that,the PP algorithm had high tracking accuracy under low and medium-speed conditions,and weak stability and low tracking accuracy under high-speed conditions;the MPC algorithm had good lateral tracking accuracy under low,medium,and high-speed conditions,but the real-time performance was weak.Considering the amount of computation and real-time performance,PP algorithm was preferred in low-speed working conditions,and MPC algorithm was the best in high and medium-speed working conditions.