基于马氏距离的极区容错联邦滤波算法  

Polar fault-tolerant federated filtering algorithm based on mahalanobis distance

作　　者：程建华 周欣进 刘佳鑫 常乐 CHENG Jian-hua;ZHOU Xin-jin;LIU Jia-xin;CHANG Yue(College of Intelligent Systems Science and Engineering,Harbin Engineering University,Harbin 150001,china)

机构地区：[1]哈尔滨工程大学智能科学与工程学院,黑龙江哈尔滨150001

出　　处：《舰船科学技术》2024年第8期143-148,共6页Ship Science and Technology

基　　金：国家自然科学基金资助项目(62073093,62003108)。

摘　　要：针对舰船组合导航系统受极区极端恶劣环境影响发生外部量测故障从而降低系统精度的问题,提出基于马氏距离的极区容错联邦滤波算法。通过引入真实量测与预测量测之间的马氏距离来衡量各个子滤波器的量测质量,并通过马氏距离来构建各个子滤波器在主滤波器最优融合中对应的融合系数,降低故障子滤波器在最优融合中的比例,减少故障子滤波器对全局估计的不良影响。实验仿真表明,该算法相较于传统的联邦滤波算法,尤其在子滤波器量测故障情况下,具有更好的容错性和可靠性。Aiming at the problem that the integrated navigation system of ships suffers from external measurement failure due to the extreme harsh environment in the polar region,which reduces the system accuracy,a polar fault-tolerant federated filtering algorithm based on Markov distance is proposed.The measurement quality of each sub-filter is measured by introducing the mahalanobis distance between the real measurement and the predicted measurement,and the corresponding fusion coefficient of each sub-filter in the optimal fusion of the main filter is constructed by using the mahalanobis distance,so as to reduce the ratio of the fault sub-filter in the optimal fusion and reduce the adverse effect of the fault sub-filter on the global estimation.The simulation results show that this algorithm has better fault-tolerance and reliability than the traditional federated filtering algorithm,especially in the case of sub-filter measurement failure.

关 键 词：联邦滤波 极区组合导航 量测故障 信息融合 

分 类 号：U666.1[交通运输工程—船舶及航道工程]