分布式驱动电动汽车的平方根容积卡尔曼滤波状态观测

State observation of distributed drive electric vehicle using square root cubature Kalman filter

针对车辆动力学系统状态估计的非线性问题,引入非线性动态Dugoff轮胎模型来构建包括纵向、侧向、横摆和侧倾等8自由度的非线性车辆动力学状态估计系统.在融合车载多传感器信息的基础上设计了车辆动力学的平方根容积卡尔曼非线性滤波状态观测器,对质心侧偏角、轮胎侧向力等关键状态进行观测.在Matlab/Simulink环境中搭建了Simulink-Carsim分布式驱动电动汽车系统状态估计联合仿真平台,采用双移线工况对观测器的可行性和有效性进行仿真验证.结果表明:传统的扩展式卡尔曼滤波状态观测器在车辆经历高侧向加速度过程中的观测值大幅偏离车辆运行状态的真实值,而设计的平方根容积卡尔曼非线性滤波状态观测器在整个双移线仿真工况下观测结果平稳,能实时反映车辆动力学系统的真实非线性运行状态,具有更小的观测误差和更高的观测精度.

To deal with nonlinear challenges on vehicle dynamics state estimation,the eight-DOF（ degree of freedom） nonlinear vehicle dynamics state estimation system,including longitudinal,lateral,yaw,and roll motions was constructed by introducing a nonlinear dynamics Dugoff tire model.Based on multi-sensor data fusion,the nonlinear observer with square root cubature Kalman filter was designed to estimate some key parameters,such as lateral tire-road forces and vehicle sideslip angle. Then the co-simulation platform with Simulink-Carsim for the estimated system of distributed drive electric vehicles was built in Matlab / Simulink environment. Simulations for double lane change manoeuvre were carried out to evaluate the feasibility and the effectiveness of the observer. The results showthat the observed values with traditional extended Kalman filter state observer deviate from the real values of the vehicle running state when vehicles deliver high lateral acceleration,while the nonlinear observer with the proposed square root cubature Kalman filter has smooth results and reflects the real-time nonlinear vehicle dynamics state during double lane change manoeuvre. And it possesses smaller observer errors and higher observation precision.