一种基于Huber法改进的Sage-Husa自适应算法  

A Sage-Husa Adaptive Algorithm Improved by Huber Method

作　　者：高帅 梁凯 GAO Shuai;LIANG Kai(Yunnan Minzu University,Kunming 650031)

机构地区：[1]云南民族大学,昆明650031

出　　处：《舰船电子工程》2025年第3期144-147,共4页Ship Electronic Engineering

摘　　要：由于GNSS和INS两者性能有很强的互补性,为了得到最佳的导航效果,组合导航的设计通常采用卡尔曼滤波器来处理各种数据,当发生量测异常时,不能解决卡尔曼滤波和Sage-Husa自适应滤波的滤波发散和实时预估状态测量噪声阵的问题。针对以上问题,提出一种改进的自适应滤波,通过在检测到测量异常时实时对新息进行判别,筛选可用的量测信息,通过改进的Huber算法对自适应滤波的观测方程进行重新构建,进而对状态估计均方误差阵进行优化,有效解决组合导航中测量异常和滤波发散的问题。仿真证明:对比Sage-Husa自适应滤波算法,改进后的自适应滤波提高了上述算法的稳定性与可靠性,改善了组合导航系统的性能。Because GNSS and INS have strong complementary performance,in order to obtain the best navigation effect,the design of integrated navigation usually uses Kalman filter to process various data.When measurement anomaly occurs,it cannot solve the problems of filter divergence and real-time estimation of state measurement noise matrix of Kalman filter and Sage-Husa adaptive filter.Aiming at the above problems,an improved adaptive filter is proposed,which discriminates the innovation in real time when detecting measurement anomaly,filters the available measurement information,reconstructs the observation equation of adaptive filter by improved Huber algorithm,and then optimizes the state estimation mean square error matrix,effectively solving the problems of measurement anomaly and filter divergence in integrated navigation.Simulation proves that compared with Sage-Hu-sa adaptive filtering algorithm,the improved adaptive filtering improves the stability and reliability of the above algorithm and im-proves the performance of integrated navigation system.

关 键 词：组合导航 Huber法 自适应滤波 鲁棒性 

分 类 号：TN967.2[电子电信—信号与信息处理] TP391[电子电信—信息与通信工程]