基于弯道自适应强跟踪卡尔曼滤波的侧向坡度估计算法

Road Bank Estimation Based on Curve Adaptive Strong Tracking Kalman Filter

摘要:

道路的侧向坡度直接影响车辆侧向运动，侧向坡度估计已成为智能汽车稳定控制系统的关键部分之一。然而，侧向坡度与车身侧倾之间存在耦合，且侧向力估计困难，准确的侧向坡度估计难度较大。为此，提出了一种基于加速度传感器的可拓融合侧向坡度估计算法：首先，提出加速度传感器模型和车辆侧倾模型，采用弯道自适应强跟踪卡尔曼滤波算法（CASTKF）对侧向坡度进行估计；然后，提出基于侧向加速度传感器的直接估计方法，防止CASTKF算法在失去可观性后的错误估计；再后，利用可拓算法对两种模式的估计值进行数据融合；最后，采用硬件在环测试（HIL）验证所提算法的有效性。结果表明，智能汽车的侧向坡度估计中采用CASTKF融合算法具有更高的精度和鲁棒性。

Abstract:

The road bank angle directly affects the lateral dynamics of the vehicle. Bank angle has become one of the key parameters of the intelligent vehicle stability control system. However， not only the coupling problem between road bank and vehicle roll， but the difficulty of getting lateral force makes the accurate estimation of bank angle a challenging problem. Therefore， an extension fusion road bank estimation algorithm based on the acceleration sensor was proposed in this paper. First， a lateral acceleration sensor model and the roll dynamics model were proposed， and the curve adaptive strong tracking Kalman filter （CASTKF） was used to estimate the road bank. Then， a direct estimation method based on the lateral acceleration sensor was proposed to prevent the wrong estimation after the loss of observability of the system. Next， the extension algorithm was used to fuse the estimated values of the two methods. Finally， hardware-in-loop tests （HIL） were conducted to verify the effectiveness of the proposed algorithm under various working conditions and the results revealed the accuracy and robustness of the CASTKF algorithm.