摘要
针对无人水下航行器(UUV)目标跟踪精度不高的问题,文中将一种鲁棒性较强的M极大似然估计代价函数引入Huber-容积卡尔曼滤波(H-CKF)并应用于UUV的目标跟踪定位算法中,通过改变归一化新息协方差的方法对CKF矩阵进行线性化求解。建立了UUV运动模型及观测模型,在不同的非高斯噪声干扰下与转换测量卡尔曼滤波、CKF和扩展卡尔曼滤波3种滤波算法进行对比试验,验证了HM-CKF的滤波精度和稳定性优于传统算法。
To improve the target tracking accuracy for unmanned undersea vehicle(UUV),a robust M maximum likeli-hood estimation cost function is introduced into Huber-cubature Kalman filter(H-CKF)for UUV’s target tracking,and the CKF matrix is linearized by changing the normalized innovation covariance.UUV motion model and observation model are established to compare the Huber M-cubature Kalman filter(HM-CKF)with the converted measurement Kal-man filter,the cubature Kalman filter and the extended Kalman filter(EKF)under different non-Gaussian noise interfer-ences,and the results show higher filtering precision and stability of the HM-CKF than the traditonal algorithm in com-plicated undersea acoustic environment.
作者
王斌
温泉
范世东
WANG Bin;WEN Quan;FAN Shi-dong(School of Energy and Power Engineering,Wuhan University of Technology,Wuhan 430063,China;Changjiang Sea-route Planning Design Research Institute,Wuhan 430010,China)
出处
《水下无人系统学报》
北大核心
2020年第1期39-45,共7页
Journal of Unmanned Undersea Systems
基金
中央高校基本科研业务费专项资助项目(195205013)
关键词
无人水下航行器
卡尔曼滤波
M极大似然估计代价函数
unmanned undersea vehicle(UUV)
Kalman filter
M maximum likelihood estimation cost function
