Identification Method of SINS Error Parameters Utilizing Artificial Neural Network for Hypersonic Vehicle

To improve the navigation accuracy of hypersonic vehicle, an error parameter identification method of strap-down inertial navigation system (SINS) based on artificial neural network is proposed. Firstly, the inertial measurement unit (IMU) error model and the SINS navigation calculation model are established, which can provide an accurate model basis for the error parameters identification. Then, four kinds of neural network structures with different inputs are constructed and optimized by the numerical simulation. To further improve the identification accuracy of the error parameters, the neural network method is improved by adjusting the inputs of the network structure and optimizing the initial network values based on gradient particle swarm optimization. The improved neural network method is eventually used to identify the SINS error parameters of hypersonic vehicle. The simulation results show that the neural network with position deviation, velocity deviation and attitude deviation as the inputs has the optimal network structure. The improved artificial neural network method has the highest identification accuracy for the error parameters compared with other methods, and the relative errors of the error parameters are less than 34%. Meanwhile, the navigation errors of the proposed method are equivalent to the magnitude of Kalman filter algorithm, which demonstrates the effectiveness of our method to identify the high-precision SINS error parameters and improve the navigation accuracy for hypersonic vehicle.