基于支持向量机的车辆质心侧偏角估计

Vehicle sideslip angle estimation based on support vector machine

针对动力学建模方法对车辆质心侧偏角进行估计所面临的路面附着系数和车辆参数无法准确获取等缺点,基于统计学理论中的支持向量机对车辆质心侧偏角估计展开研究。选择方向盘转角、车辆速度、横摆角速度和侧向加速度作为支持向量机的特征向量。在Carsim仿真平台设计了20组典型车辆操纵试验作为训练样本得到预测模型,通过2组变附着系数路面上的操稳性试验对模型进行了验证。研究结果表明:支持向量机可以有效实现对不同附着路面上车辆质心侧偏角的估计,达到了较高的估计精度,即使车辆发生大侧偏现象使轮胎进入侧偏角-侧偏力曲线的非线性域,该方法仍能够实现质心侧偏角的准确估计,估计的绝对误差不超过1.42°,从而为车辆主动安全控制提供了参考。

For dynamic model-based method to estimate the vehicle sideslip angte, tne alsauvanta -ges such as the uncertain vehicle parameters and unknown road friction coefficient are obvious. In this paper, the support vector machine （SVM） was proposed to realize the estimation of the vehi- cle sideslip angle based on statistics theory. The characteristic vectors for SVM were chosen in- cluding the steering angle, the vehicle speed, the yaw rate and the lateral acceleration. The train- ing samples were obtained on the Carsim simulation platform in which a group of 20 maneuver tests were conducted to establish prediction model. The model was verified by two groups of ve- hicle controllability and stability tests on variable adhesion coefficient road. The results show that the method can realize the accurate estimation of the sideslip angle and the absolute error does not exceed I. 42° even when the severe sideslip occurs and the tire takes on the nonlinear field, which can provide useful reference for the active safe control. 1 tab, 14 figs, 11 refs.