纵横向运动耦合时车辆状态估计算法研究

Research on Vehicle State Estimation Based on Longitudinal Lateral Motion Coupling

考虑到车辆纵向运动和横向运动的主要耦合因素,提出了一种利用扩展卡尔曼滤波理论间接测量车辆行驶状态参数的方法。首先,研究了卡尔曼滤波理论及其算法的具体流程,建立了基于Dugoff轮胎模型的耦合三自由度动力学模型,结合基于横向加速度反馈的预瞄最优曲率驾驶员模型,建立了"人-车"闭环整车系统;其次,搭建了基于扩展卡尔曼滤波理论的车辆行驶状态估计器仿真平台;最后,对某车型给定蛇形路径行驶工况进行了仿真。结果表明:该驾驶员模型能很好地跟踪车辆的横向轨迹,且前轮转向适当,易于实现;借助车辆易测得的纵向、横向加速度信息,结合扩展卡尔曼滤波算法能准确地估计运动耦合条件下车辆的纵向速度、横向速度和横摆角速度,且误差控制在5%以内。

Considering the main coupling factors of the vertical and horizontal directions of the vehicle,a method of indirect measurement of the vehicle driving state parameters by using the extended kalman filter theory is proposed.Firstly,the kalman filter theory and its algorithm were studied.The coupled three-degree-of-freedom dynamic model based on Dugoff tire model was established.Combined with the lateral acceleration feedback-based preview optimal curvature driver model,a"human-car"closed-loop vehicle system was established;Secondly,the vehicle driving state estimator simulation platform based on extended kalman filter theory was built.Finally,the driving situation of a given snake-shaped path was simulated.The results show that the driver′s model can track the lateral trajectory of the vehicle well,and the front wheel steering is appropriate and easy to implement.With the easy-to-measure longitudinal and lateral acceleration information of the vehicle,combined with the extended kalman filter algorithm,the longitudinal velocity,lateral velocity and yaw rate of the vehicle under motion coupling conditions can be accurately estimated,and the error is controlled within 5%.