非充分激励条件下路面附着系数估计算法

Tire-road friction coefficient estimation algorithm under insufficient excitation conditions

针对路面条件和车辆状态激励程度的不确定性导致的路面附着系数算法收敛速度和估计精度下降的问题,提出了一种基于模糊工况自适应强跟踪卡尔曼滤波的路面附着系数估计算法。利用模糊推理方法评估当前车辆运动状态的激励程度并输出协方差调整系数,引入强跟踪因子对标准卡尔曼滤波算法进行实时修正,通过及时调整路面附着系数的协方差的方式提高估计算法收敛速度,同时强跟踪因子保证算法对来自路面不确定的扰动具有鲁棒性。采用控制器硬件在环试验台的方式对所提算法的估计效果进行了验证,实验结果表明:所提出估计方法能够在车辆状态大激励程度条件时快速收敛到真值附近,小激励程度时降低估计值波动幅值,比强跟踪卡尔曼滤波算法和标准卡尔曼滤波算法在算法收敛速度和估计精度方面有明显提升。

The convergence speed and estimation accuracy of the pavement adhesion coefficient algorithm are reduced due to the uncertainty of road condition and vehicle state excitation.This paper presents a road adhesion coefficient estimation algorithm based on adaptive strong tracking Kalman filter under fuzzy operating conditions.The fuzzy inference method is used to evaluate the excitation degree of the current vehicle state and output the covariance adjustment factor.A strong tracking factor is introduced to correct the Kalman filter algorithm in real time.By adjusting the covariance of the road adhesion factor,the convergence speed of the estimation algorithm is improved,and the strong tracking factor ensures that the algorithm is robust to disturbances from the road surface uncertainty.The estimation effect of the proposed algorithm is validated by a hardware-in-loop test bench.The experimental results show that the proposed estimation method can quickly converge near the true value under large excitation conditions and reduce the amplitude of the fluctuation of the estimated value under small excitation conditions.Compared with strong tracking KF algorithm and KF algorithm,the proposed algorithm markedly improves the algorithm convergence speed and estimation accuracy.