| Navigation sensor failure in the INS/GPS integrated navigation system may impact the integrity of the system in highly dynamic condition severely. The fault data of navigation sensor directly used in filtering process will make the navigation results aberration, and even cause serious results. In this paper, I choose the INS/GPS integrated navigation system as the researching object, then I have read a large number of literatures about particle filter and integrated navigation fault diagnosis and done some researches about data fusion and fault diagnosis about integrated navigation, and the paper mainly includes the following aspects:1. This proposed a UPF filter estimation scheme for INS/GPS integrated navigation system based on its loosely coupled model. In the paper, navigation parameters and error angle of INS platform were used as the system states, the combination of mechanical arrangement equations of inertial system and attitude error equation was constructed as the state equation, the output of GPS was used as the measurement. and then I used UPF to calculate the parameters of integrated navigation system directly.2. According to the characteristics of INS/GPS integrated navigation system, I proposed a fault diagnosis method and a hierarchical fault diagnosis method for integrated navigation system based on UPF. Then a feasible fault diagnosis scheme with system residual generated through multiple filter was carried out, and the fault isolation process was completed by using GLRT(Generalized Likelihood Ratio Test) method.3. In order to improve the particle’s diversity of particle filter algorithm, an improved particle filter algorithm was proposed. In order to make the particle deviates from its equilibrium position, after the completion of multinomial sampling, I use the Gauss threshold regulator disturbance into progeny particle set, this method can improve the convergence speed and robustness of the particle filter algorithm. |
PDF Full Text Request