基于意图识别算法的转矩控制策略研究

Research on Torque Control Strategy Based on Intention Recognition Algorithm

通过Adaboost算法进行了驾驶员意图识别的转向力矩分配,并搭建Simulink仿真模型,在Carsim协同Simulink仿真的前提下,进行了蛇形实验、最小转弯半径实验以及双移线实验等典型工况实验。实验结果表明在用Adaboost算法进行驾驶员意图识别的基础上,进行转向力矩分配具有很高的准确性,并且实验工况条件设置一样的情况下,识别驾驶意图的转矩分配策略能够给驾驶员更加舒适的转向性能,且车辆在低速掉头行驶时的通过性和高速避障的稳定性都比传统力矩分配的控制效果好。

A steering torque distribution,and to build simulation models,Carsim and Simulink cooperate simulation,experi⁃ments conducted serpentine,and the minimum turning radius double lane change test experiments described herein were based on Adaboost algorithm driver intention recognition.Verified by experiment based on the use of Adaboost algorithm to identify the driver's intention on the steering torque distribution with high accuracy,and in the same experimental conditions,using the model to predict driver intention recognition control torque distribution strategy,can provide a more supple act of the vehicle.Low-speed,wide-angle steering and high-speed driving stability and operability are better than traditional control strategies.