APP下载

基于迭代扩展卡尔曼滤波的车辆运动状态估计

2021-04-01赵强范思远唐政林

森林工程 2021年1期

赵强 范思远 唐政林

摘 要:采用低成本传感器并借助卡尔曼滤波方法实现车辆运动状态的高精度估计。首先考虑车辆侧向运动、横摆运动以及侧倾运动,建立非线性三自由度的动力学车辆模型,通过对其线性化,实现扩展卡尔曼滤波设计,进一步针对线性化带来的截断误差问题,利用贝叶斯估计建立极大后验状态估计最小二乘表达式,通过进一步求解最终设计完成了迭代扩展卡尔曼滤波算法。通过不同行驶条件下仿真,验证迭代扩展卡尔曼滤波过滤噪声和追踪实际值的能力。仿真结果表明:在复杂的行驶条件下,迭代扩展卡尔曼滤波能大幅过滤噪聲,并有效追踪车辆质心侧偏角和横摆角速度的实际状态。

关键词:车辆运动状态估计;迭代扩展卡尔曼滤波;三自由度;质心侧偏角;横摆角速度

中图分类号:U463.6    文献标识码:A   文章编号:1006-8023(2021)01-0066-07

Abstract:The paper discusses to use low-cost sensors and Kalman filter to achieve high-precision vehicle motion state estimation. Firstly, considering the lateral, yaw motion and roll motion of the vehicle, a nonlinear 3-DOF dynamic vehicle model is established and further linearized to realize extended Kalman filter design. To solve the problem of truncation error caused by linearization, the least square expression of largest posterior state estimation is established by Bayesian estimation, and the iterative extended Kalman filter (IEKF) algorithm is finally completed by further solving. Through simulations under different driving conditions, the IEKF algorithm is verified on filtering noise and tracking the actual value. The simulation results show that under complex driving conditions the IEKF algorithm can effectively filter noise and track the actual state of the vehicle slip angle and yaw rate.

Keywords:Vehicle motion state estimation; iterative extend Kalman filter; 3-DOF; slip angle; yaw rate

0 引言

实时准确获取汽车行驶状态是汽车主动安全控制的必要条件[1]。由于成本的限制,高精度但价格昂贵的状态测量传感器通常较难装车应用,实际的车辆一般采用低成本传感器结合滤波算法进行行驶状态的估计[2]。状态估计大致分为运动学估计和动力学估计两类[3-4]。

Koibuchi等[5]运用运动学估计,通过运动学模型,对传感器信号运用直接积分法估计。Hac等[6]使用卡尔曼滤波器优化汽车运动学估计模型的参数。以上结果均有一定精度,但运动学模型对参数变化缺乏实时性,对设备精度依赖程度高。因此近年来对汽车状态估计大多采用动力学模型[7-8]。……

登录APP查看全文