Research On Fault Diagnosis Of Integrated Navigation System Based On IGRMMAE Algorithm

Because inertial navigation and satellite navigation are complementary to each otherwell, the combination of inertial and satellites becomes a more desirable navigation mode.Especially in recent years, with the development and gradual application of multiple ofnew technologies in navigation, the application based on integrated navigation is more andmore extensive.There are some new challenges for integrated navigation system because of itsinherent structural complexity and it’s easily influenced by multiple environmental factors,which mainly embodied in diversity of system fault and sensitivity of system to fault.According to the fact that hard fault, soft fault, persistent fault and intermittent fault easilyoccur in the integrated navigation system under the environment of high speed and highdynamic, and the requirements for great reliability, adaptability and real time, this paperputs forward a new diagnosis algorithm for improving the real-time performance andadaptability of fault detection in integrated navigation system, namely,IGRMMAE(Interacting Generalized Residual Multiple Model Adaptive Estimator), whichimproves the rapid detection ability of integrated navigation system to fault by introducinghomogenization residual and the adaptability to various faults by adopting interactingmultiple model Kalman filters, furthermore, with the combination of homogenizationresidual and interacting multiple model Kalman filters, realizes the multi-objectivecompatibility of real-time performance and adaptability of fault detection.To verify the superiority of IGRMMAE algorithm in fault detection and diagnosis ofintegrated navigation system, this paper takes SINS/GPS/GLONASS/COMPASSintegrated navigation system for example, establishes a mathematical model and faultmodel of the integrated navigation system in MATLAB simulation platform. Thesimulation results demonstrate that this algorithm can implement fault detection andisolation, fuse the data based on the result of diagnosis, result in active fault-tolerantguidance under the influence of various time-varying faults.