Application of Wavelet Neural Network to Initial Alignment of Strapdown Inertial Navigation System

This paper describes a new initial alignment method for strapdown inertial navigation system(SINS) on stationary base. Wavelet neural network(WNN) is used in the method, which solves the problem that azimuth error has slow convergence rate in Kalman filter. The methodology is analyzed deeply and the gradient descent method is used to deduce the iterative formulas of the network parameters in detail. The simulation of the application of WNN and Kalman filtering methods to initial alignment is done separately. The simulation results show that the new method has faster convergence speed and higher precision than Kalman filtering method. It can meet the real time requirement better.