Research On INS/GPS/DVL Integrated Navigation Fault Tolerance Algorithm Based On Federal Kalman Filter

Today,science and technology are developing rapidly,and the field of navigation technology has also developed rapidly.It has been widely used in all areas of society,especially in the military field.Currently commonly used navigation systems include: Inertial Navigation System(INS),Global Positioning System(GPS),Doppler Velocity Log(DVL).The advantages of multiple navigation sensors are combined,and combined navigation can overcome the disadvantages of a single sensor.As the number of sensors is increased,the possibility of failure of the combined system will also increase greatly.Therefore,it is necessary to add a suitable fault detection link to the integrated navigation system to perform real-time fault detection and isolation of the navigation subsystem.Therefore,the fault tolerance performance of the integrated multi-sensor integrated navigation system is improved.In this paper,the fault-tolerance method of INS / GPS / DVL integrated navigation system is studied.The global satellite navigation system,Doppler log and inertial navigation system are combined respectively,and the global information fusion is implemented by using federal Kalman filtering technology.In the federal Kalman filter structure,model-based and data-driven fault detection methods are designed respectively: the model-based detection method has residual chi-square fault detection and dual-state chi-square fault detection;the data-driven detection method is the principal component Analyze fault detection and identification.Through these two kinds of fault detection methods,real-time fault detection and isolation of the sub-filter is realized,and the overall fault tolerance performance of the navigation system is improved,thereby improving the navigation accuracy of the system.First,the basic theory of the multi-sensor integrated navigation system is introduced.The theory of integrated navigation system is analyzed,and the discrete Kalman theory and the federal Kalman filter theory are derived.The working principle and error model of the inertial navigation system are analyzed,and the global satellite navigation system and Doppler log are introduced.Then,complete the modeling and simulation design of the Federal Kalman Filter system.The inertial navigation system is combined with the global satellite navigation system and the Doppler log,and the federal Kalman filter design is used,and MATLAB is used for simulation.The single combination is compared with the federated Kalman combination,and the federated Kalman with feedback and the federated Kalman without feedback are compared.In order to improve the fault tolerance performance of the system,model-based fault detection methods are added to the federal Kalman filter structure: residual chi-square and dual-state chi-square fault detection,and simulation verification is performed.Finally,a fault detection and identification method based on principal component analysis is added to the federal Kalman filter structure.The working process of fault detection based on principal component analysis is analyzed: the data of the navigation system under normal conditions is used to establish the principal component model,and then the system performs fault detection online.After a fault occurs,fault identification based on the contribution graph method is used to realize accurately locate the source of the fault,and verify the feasibility of the method through simulation.