Error estimation of airborne strapdown inertial navigation system based on neural network

Abstract

Strapdown Inertial Navigation System play an important role in many different kinds of applications. As the accuracy of which is deeply influenced by the sensors precision, so the system error propagation should be addressed. Based on the airborne vehicle characteristics, a method based on neural network is proposed to estimate the attitude, velocity and position error of system. The simulation experiment results validate the algorithm in estimate some kind of system errors is better than the traditional method based on Kalman filter model.