中国空间科学技术

• •    

基于递阶联合滤波器的故障检测与重构方法

董锡君,王宇飞,洪兴昌   

  1. 哈尔滨工业大学航天学院!哈尔滨150001,北京京惠达新技术公司!北京100086,上海航天局!上海200233
  • 发布日期:2001-04-25

Fault Detection and Recovery Method Based on Hierarchic Alliance of Filters

Dong Xijun (School of Astronautics,Harbin Institute of Technology,Harbin 150001) Wang Yufei (Beijing JHD New tech Company,Beijing 100086) Hong Xingchang (Shanghai Space Industry Bureau,Shanghai 200233)   

  • Online:2001-04-25

摘要: 随着现代航天器等高精度飞行系统的日益复杂 ,飞行控制系统的安全运行问题愈来愈突出地表现出来。故障检测、隔离和重构 ( FDIR)技术是航天器可靠性的重要保证。文章在卡尔曼滤波估计算法的基础上 ,提出了基于递阶联合滤波器族的故障检测、隔离和重构方法。该方法根据备份滤波器的概念 ,构造了一种融合滤波器的树状递阶联合结构 ,在此基础上 ,可以设计并实现多级故障检测和隔离措施 ,进而有效地完成系统的重构功能。通过一个 INS/GPS/Doppler组合导航系统的仿真实例 ,说明了该方法的可行性和有效性。基于递阶联合滤波器结构的多传感器信息处理技术克服了集中卡尔曼滤波器和普通联合滤波器的缺陷 ,对提高多传感器系统的容错性能具有重要意义

关键词: 卡尔曼滤波, 故障检测, 故障仿真, 多传感器

Abstract: As the high precision aircraft such as the modern spacecraft becomes more complicated,the safe operating problem also becomes more and more significant.FDIR (Fault Detection,Isolation and Recovery) techniques are important guarantee of the reliability of spacecraft.Based on Kalman filtering estimation algorithm,a kind of FDIR method based on hierarchic alliance of filter family is proposed.According to the concept of the backup filter,this method constructs a kind of dendriform hierarchic alliance structure of fusion filter.On the basis of that,the multilevel FDIR can be designed and performed.A simulation example of INS/GPS/Doppler integrated navigation system illustrates the feasibility and effectiveness of this method.The multi sensor information processing technique based on the hierarchic alliance filter structure can resolve the limitation of the centralized Kalman filter and the normal federated filter,and has significance to enhance the fault to lerant ability of the system.Fault Detection,Isolation and Recovery) tech