基于修正量测的UKF算法  

UKF Algorithm Based on Modified Measurement

作　　者：崔灿 刘攀 张敏[2] CUI Can;LIU Pan;ZHANG Min(School of Mechanical Engineering,Hubei Univ.of Tech.,Wuhan 430068,China;2 Engineering and Technology College.Hubei Univ.of Tech.,Wuhan 430068,China)

机构地区：[1]湖北工业大学机械工程学院,湖北武汉430068 [2]湖北工业大学工程技术学院,湖北武汉430068

出　　处：《湖北工业大学学报》2021年第4期4-7,共4页Journal of Hubei University of Technology

摘　　要：针对Kalman滤波(UKF)不能解决非线性算法的难点,在Kalman滤波的算法上加以改进。由于(UKF)算法需要利用无迹变换(UT)求解非奇异协方差矩阵的平方根,但在求解奇异协方差矩阵或滤波计算时会有较大的误差,导致算法的精度无法保证。提出了一种基于修正测量UT变换的修正测量UKF算法来处理奇异协方差矩阵来解决这一问题,,并通过仿真验证了该方法的有效性。Directed at the difficulty that Kalman Filtering cannot solve the pain points of non-linear algorithms,an Unscented Kalman Filter(UKF)is proposed.This algorithm needs to use the unscented transformation(UT)to find the square root of a non-singular covariance matrix,but it is solving Or a large error will occur in the filtering calculation,and the accuracy of the algorithm cannot be guaranteed.In order to analyze this problem,a UT transform based on modified measurement is proposed to deal with the singular covariance matrix,which constitutes the modified measurement UKF algorithm,and the effectiveness of this method is proved by simulation.

关 键 词：无迹滤波 UT变换 修正量测 奇异矩阵 

分 类 号：TN911[电子电信—通信与信息系统]