基于EKF对汽车质心侧偏角的估计

Estimation of Sideslip Angle Based on Extended Kalman Filter

通过道路实车试验,验证了Car Sim的仿真模型的准确性,为后续的估计提供仿真实验平台。用Car Sim代替实车,基于卡尔曼滤波KF(Kalman filter)用二自由度车辆模型;基于扩展卡尔曼滤波EKF(Extended Kalman Filter)用三自由度车辆模型,分别在低速和高速两种工况下,建立Car Sim和MATLAB/Simulink联合仿真,估计质心侧偏角,比较估计效果。实验结果表明:由于汽车在低速行驶的工况下车辆动力学特性大致呈线性变化,用KF和EKF估计值差别不大,但是汽车在高速行驶的工况下车辆动力学特性趋于非线性,KF估计结果误差较大,而EKF估计的适应性要好一些。

The accuracy of CarSim simulation model is verified by road test, and the simulation experiment platform is provided for subsequent estimation. Using CarSim replacing real vehicle, vehicle 2-DOF model based on Kalman filter （KF） and 3-DOF model of vehicle based on extended Kalman filter （EKF） are used at low speed and high speed. The co-simulation model of CarSim and MATLAB/Simulink was created to estimate the sideslip angle and compare estimated results. The experimental results show that the dynamic characteristics of the vehicle are almost linearly changed in low speed driving conditions, and the estimated value of KF and EKF is not very different, but the vehicle＇s dynamic characteristics tend to be nonlinear at high speed. KF estimates the results of a larger deviation, and the adaptability that EKF estimates is better.