Information Fusion And Fault Tolerance Theory Of Intelligent Integrated Navigation System

A modern aircraft with high performance requires that its navigation system have high accuracy and high reliability. It's impossible to meet both demands by any kind of navigation equipment separately. Nowadays integrated navigation system based on Kalman Filtering theory has efficiently increased the cost performance ratio. But when the classical Centralized Kalman Filter (CKF) is used in this kind of integration, there will be two disadvantages: firstly, there exists real time computation difficulty; secondly, the fault tolerance capability is poor. Once one of the subsystems breaks down, all the other subsystems are affected. The Federated Kalman filter raised by Carlson, which is developed from the decentralized filter, can be used to design the integrated navigation system with high precision and good fault tolerance flexibly.In this thesis, on the basis of Federated Kalman filtering theory and aiming at a certain type of airplane equipped with many navigation equipments such as SINS, GPS,etc., an integrated navigation system with the fault tolerance property is designed. The thesis is stressed to investigate some new theories and methods of FDI in designing the fault tolerance integrated navigation system. The following important work has been done via simulations: ?The performance of the federated filter designed is compared with that of CKF; ?The component level fault detection of SINS is investigated. Generalized Likelihood Ratio Test Method is introduced. Fault detection method and simulation results are presented when the number of fault gyro is one or two. (3) This is my innovation of the thesis: a new method-adaptive filtering algorithm is proposed to estimate the combinations of sensor errors affecting the parity vector, then the estimates are used to form a compensated parity vector. Compared with the conventional separated-bias estimation method, parity vector compensation effect has been simulated. ?The State x2 Test Method and the Residue Test Method have been developed in the system level fault detection of the local filter (LF). Directing toward the SINS/GPS system, the fault detection simulation is carried out with the adaptive filtering algorithm, and a good result is achieved.