• DocumentCode
    133691
  • Title

    Fault detection and diagnosis in the INS/GPS navigation system

  • Author

    Wen Xin ; Ji Long ; Zhang Xingwang ; Zhao Jianxin

  • Author_Institution
    Coll. of Aerosp., Shenyang Aerosp. Univ., Shenyang, China
  • fYear
    2014
  • fDate
    3-7 Aug. 2014
  • Firstpage
    27
  • Lastpage
    32
  • Abstract
    This paper proposed a new fault detection and diagnosis (FDD) scheme based on INS (inertial navigation system) /GPS integrated navigation system. Once a fault is detected in the process of fault diagnosis, we should complete two diagnosis phases: primary diagnosis and diagnosis and isolation phase. In the primary diagnosis phase, we use filter to generate residuals that we need, then complete the fault detect based on chi-square test of residuals; In the isolation phase, we design a multiple hypothesis test for innovation sequence of the navigation system, then complete fault isolation based on generalized likelihood ratio(GLR) test. This scheme can reduce the difficulty and complexity of fault diagnosis, improve the diagnosis accuracy, and can identify the concurrent fault, strengthen anti interference ability. In addition, even if only one navigation device (INS or GPS) works properly, this scheme can still guarantee system to achieve a sufficiently accurate state estimation.
  • Keywords
    Global Positioning System; fault diagnosis; inertial navigation; interference suppression; INS-GPS integrated navigation system; chi-square test; fault detection; fault diagnosis; fault isolation; generalized likelihood ratio test; inertial navigation system; innovation sequence; isolation phase; Accelerometers; Fault detection; Fault diagnosis; Global Positioning System; Silicon compounds; Vectors; Fault diagnosis; GLR test; INS/GPS integrated navigation system;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    World Automation Congress (WAC), 2014
  • Conference_Location
    Waikoloa, HI
  • Type

    conf

  • DOI
    10.1109/WAC.2014.6935645
  • Filename
    6935645