%0 Journal Article
%T 采用自适应无迹卡尔曼滤波器的车速和路面附着系数估计<br>Estimation of Vehicle Speed and Tire？？Road Adhesion Coefficient by Adaptive Unscented Kalman Filter
%A 张家旭
%A 李静
%J 西安交通大学学报
%D 2016
%R 10.7652/xjtuxb201603011
%X 针对车辆主动安全控制中的车速和路面附着系数这一关键信息,提出了一种实时估计该信息的滤波算法,同时建立了将包含时变噪声统计特性的七自由度非线性车辆动力学模型作为滤波算法的标称模型,以及一种自适应无迹卡尔曼滤波算法。该算法采用传统的无迹卡尔曼滤波器来估计车速和路面附着系数,同时利用次优Sage-Husa噪声估计器对系统的噪声统计特性进行实时更新,其中采用遗忘因子限制噪声估计器的记忆长度,使新近数据发挥重要作用,使陈旧数据逐渐被遗忘,从而解决了因系统标称模型误差、外界扰动等因素引起的噪声时变的问题。在不同路面条件下进行了多种工况的实验验证,并与无迹卡尔曼滤波器的估计结果进行对比分析,结果表明,该算法具有良好的鲁棒性,其估计精度高于无迹卡尔曼滤波器,且满足车辆主动安全控制系统的要求。<br>An adaptive unscented Kalman filter (AUKF) algorithm for estimating vehicle speed and tire？？road adhesion coefficient, the essential information for the active safety systems, is proposed. 7-DoF nonlinear vehicle dynamics model containing varying statistical noise characteristics is established as the nominal model. To solve the effects from varying statistical noise characteristics on the estimation accuracy and stability, the proposed algorithm adopts the traditional unscented Kalman filter to estimate vehicle speed and tire？？road adhesion coefficient, and the suboptimal Sage？？Husa noise estimator is used to update the statistical noise characteristics of the system simultaneously, where the forgetting factor limits the memory length of noise estimator to enhance the role of the new data and to forget the old data gradually. In the real vehicle experiment environment, the performance of the proposed algorithm is verified and compared with that of unscented Kalman filter for a variety of maneuvers and road conditions. The tests indicate the better robustness and estimation accuracy of this AUKF algorithm, which meets the requirements of the active safety systems
%K 车辆动力学
%K 自适应滤波
%K 无迹卡尔曼滤波
%K 次优Sage？？Husa噪声估计器&lt
%K br&gt
%K vehicle dynamics
%K adaptive filter
%K unscented Kalman filter
%K suboptimal Sage？？Husa noise estimator
%U http://zkxb.xjtu.edu.cn/oa/DArticle.aspx?type=view&id=201603011