Research On Data Fusion Technologies Of Projectile-borne INS/GNSS Integrated Navigation

Precision-guided artillery has attracted wide attention due to its low cost-efficiency ratio.The development of Inertial Navigation System/Global Navigation Satellite System(INS/GNSS)integrated navigation technology has further improved the strike accuracy of guided artillery.In order to further improve the precision of projectile-borne INS/GNSS integrated navigation system and its anti-jamming ability in complex battlefield environment,this paper takes a national defense key project as the background and takes the projectile-borne INS/GNSS integrated navigation system as the research object.The research contents include inertial device error compensation technology of projectile-borne inertial devices,high-precision integrated navigation algorithm,and satellite signal interference detection technology.The main research contents are as follows:(1)In the projectile-borne integrated navigation system,the use of Micro Electro Mechanical System-Inertial Measurement Unit(MEMS-IMU)can better achieve overload protection,but the accuracy of MEMS gyroscopes is lower than that of laser and fiber optic gyroscopes.In order to improve the measurement accuracy of the gyroscope,the error parameters of the gyroscope are identified by using the Allan variance analysis.Then the Autoregressive Moving Average(ARMA)model is used to compensate the gyroscope error,and an improved ARMA error compensation algorithm based on Adaptive Kalman Filter is proposed to further improve the compensation accuracy.combined with the timing characteristics of gyroscope data,an error compensation algorithm based on Long Short Term Memory Recurrent Neural Network(LSTM-RNN)is proposed,and the Gate Recurrent Unit(GRU)structure is used to optimize the training process of the neural network.The experimental results based on real data of MEMS-IMU show that the above algorithms can improve the measurement accuracy of MEMS gyroscope,and LSTM-RNN algorithm has the best compensation effect.(2)In order to improve the accuracy and reliability of projectile-borne integrated navigation system under high overload,a multi-receiver INS/GNSS loose integrated navigation model is designed.In the multi-receiver loose integrated navigation model,two data fusion methods,centralized and decentralized,are proposed.The centralized method directly makes the difference between the position and speed information from multiple receivers and the inertial navigation system information to construct a measurement matrix for centralized Processing.The decentralized method is based on the federal filtering architecture,and each receiver is used as a "sensor" for decentralized processing.Then the advantages and disadvantages of the two methods are analyzed.In order to make full use of the historical information of the trajectory,a multi-receiver loose integrated navigation optimization algorithm based on graph optimization is proposed to further improve the system accuracy.The experimental results show that the multi-receiver loose integrated navigation model can improve the accuracy of navigation and the reliability of the system,and the graph optimization algorithm can further reduce the navigation error of the system.(3)In order to further improve the accuracy of the integrated navigation system,a multi-receiver INS/GNSS tightly integrated navigation model is designed.In order to solve the problem of large amount of computation caused by the centralized processing method in the multi-receiver tightly integrated navigation model,a difference algorithm of measurement information is proposed to eliminate the clock-related variables in the state vector and reduce the computation.Aiming at the Kalman Filter ignoring historical state information,a tightly integrated navigation optimization algorithm based on graph optimization is proposed.The experimental results verify that the multi-receiver tight integrated navigation model can further improve the accuracy of navigation,and the graph optimization algorithm is still effective in the multi-receiver tightly integrated navigation.(4)Aiming at the problem that satellite navigation receivers are easy to be interfered and spooked in battlefield environment,two spoofing detection algorithms based on baseline statistical characteristics and position correlation of multi-receiver are proposed.In the algorithm based on the statistical characteristics of baselines,the hypothesis testing models of single baseline,dual baselines,and multiple baselines are established.The relationship functions of detection probability and omission probability of corresponding models are derived.The multi-baseline superposition and maximum/minimum detection algorithms are proposed.The Receiver Operating Characteristic(ROC)curves of different algorithms verify that the increase of baseline length and number can improve the detection performance,and the multi-baseline maximum and minimum algorithm has the best detection performance.In the position correlation algorithm,the position correlation matrix of multi-receiver is derived,the detection and determination factors are defined,and the method of eliminating outliers is given.Hardware-in-the-loop simulation of four receivers verifies the effectiveness of the algorithm.