Research And Implementation Of Integrated Inertial Navigation System Based On Data Fusion Filtering

With the rapid development of radio and satellite navigation in the marine field,GPS/INS navigation technology,integrated by Inertial Navigation System(INS)and Global Positioning System(GPS),is the focal point of research in the field of high accuracy and reliability of navigation systems.However,how to ensure the high accuracy and high noise immunity of ship navigation systems remains to be solved due to the complex operating environment at sea.In this paper,we establish the simulation application model of the integrated navigation system through simplifying the real environment carrier based on the INS/GPS integrated navigation system.Meanwhile,the error characteristics of the simulation model is analyzed in the real environment.According to multiple error sources,the data fusion filtering algorithm of the integrated inertial navigation system is studied.As a result,the improved SAGE-HUSA adaptive Kalman filter algorithm and the robust filter algorithm based on partial least squares are proposed to improve the performance of the integrated navigation system.The main research work includes:(1)Understanding the modeling process of the inertial navigation system.According to the complexity and weak stability of the inertial navigation system,it is necessary to simplify the model based on the actual model.The specific simulation model of the SINS accelerometer and gyro components is established and used to predict the trajectory of the carrier.Then,the universality and accuracy of the model built in this section are validated by comparing simulation data with the actual output data,which provides an inertial mathematics platform for the next step in the research of the integrated navigation system and provides data source for further data fusion filtering algorithms research.(2)Taking into account that the integrated navigation will bring a variety of errors in the actual environment,the error propagation characteristics of the inertial navigation during operation is analyzed.In order to better understand the error characteristics in the actual environment,the simulated operating environment of the integrated navigation system is built to realize the simulation update of the position,speed,acceleration and other information of the integrated navigation.Moreover,the simulation results of error characteristics of inertial navigation components are comparative analysis to provide the theoretical basis for establishing the optimal error model of the integrated navigation system and put forward requirements for filter design.(3)Studying the data fusion filtering algorithm based on the analysis of error propagation characteristics through integrating the two optimal error models.On the one hand,an improved adaptive Kalman filter algorithm is designed to filter the error,and the SAGE-HUSA algorithm is used to modify the model parameters in real time,which can simplify the calculation process and improve navigation system accuracy and real-time performance at the same time.On the other hand,a robust Kalman filter algorithm based on partial least-squares method is designed.Then the iterative method is used to continuously adjust the outliers in the integrated navigation.So that it can improve the anti-interference performance of the integrated navigation system.