Information Fusion And Fault Detection Based On Neural Network For Integrated Navigation System

With the modernization of war, integrated navigation system has already been indispensable to the military carriers and armed equipment. The navigation precision and reliability of integrated navigation system have a direct effect on achievement of operations and missions. Information fusion theory and fault detection technologies are important approaches to improve the precision and reliability of integrated navigation system. Neural network (NN) has a lot of excellent abilities just as parallel processing, self-adaptability and self-learning. More and more people paid attention to NN in this ten years. They considered that NN will have vast applying foreground.The method mentioned in this paper enhanced the precision and reliability of the system with combining NN and traditional methods of information fusion and fault detection.The main achievements contained in this dissertation are summarized as follows. (1). The federated Kalman filter is investigated. Incorporating a Chaotic BP neural network into the federated Kalman filter configuration for the SINS/GPS/ADS integrated navigation system is proposed. The simulation analysis for two information fusion methods SINS/GPS/ADS integrated navigation system is implemented. (2). The Fuzzy ARTMAP NN was investigated. The simulation analysis fordifferent inspection scale Fuzzy ARTMAP NN capability is implemented. (3). The method mentioned in this paper ameliorated the double state Chi-squareTest, enhanced the fault detection ability.(4). The method of fault detection base on Fuzzy ARTMAP NN is proposed. The simulation for the fault detection SINS/GPS/ADS integrated navigation system is implemented.