基于多重次渐消因子的强跟踪UKF姿态估计  

Attitude estimation of strong tracking UKF based on multiple fading factors

作　　者：钱华明[1] 黄蔚[1] 孙龙[1] 徐健雄[2] 葛磊[1] 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001 [2]电子科技大学光电信息学院,四川成都611731

出　　处：《系统工程与电子技术》2013年第3期580-585,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(61104036)资助课题

摘　　要：针对应用于飞行器姿态确定中的乘性扩展卡尔曼滤波(multiplicative extended Kalman filter,MEKF)存在精度低、鲁棒性差的缺点,提出了一种基于多重次渐消因子的强跟踪无迹卡尔曼滤波(multiple fading factors strong tracking unscented Kalman filter,MSTUKF)算法。该滤波算法克服了单渐消因子对多变量跟踪能力差的局限性,通过引入两个多重次渐消因子对预测误差协方差阵进行调整,使得不同的滤波通道具有不同的调节能力,保证预测误差协方差阵的对称性,从而实现滤波算法强跟踪性。仿真结果表明,MSTUKF的滤波精度和鲁棒性均明显优于MEKF,能够更好地满足工程应用对精度和鲁棒性的要求。Considering that the multiplicative extended Kalman filter （MEKF） has low accuracy and poor robustness in the application of spacecraft attitude determination, a strong tracking unscented Kalman filter algorithm based on multiple fading factors （MSTUKF） is presented. The algorithm overcomes the limitation of single fading factor for multivariable systems, and introduces two multiple fading factors to adjust the prediction error covariance, which can make the different filter channels possess different adjustment ability, and ensure the symmetry of the prediction error covariance matrix, thus realizing the strong tracking of the filtering algorithm. Simulation results show that MSTUKF is superior to MEKF in precision and robustness, and satisfies the requirements of precision and robustness which are emphasized in projects.

关 键 词：姿态确定 渐消因子 强跟踪 无迹卡尔曼滤波 鲁棒性 

分 类 号：U666.12[交通运输工程—船舶及航道工程]