A New Initial Alignment Algorithm for Strapdown Inertial Navigation System Using Sensor Output

Abstract

In this paper, a new alignment algorithm that uses simultaneously both open-loop and closed-loop scheme is designed to increase the convergence rate of the Kalman filter in the fine alignment stage. Generally, the initial alignment is divided into coarse and fine alignment. The fine alignment stage with the 10-state Kalman filter refines the initial estimate of the transformation matrix given by the coarse alignment algorithm. This paper derives a convergence theorem of the Kalman filter for analyzing a problem of the 10-state Kalman filter in the fine alignment. In order to resolve the problem, the new alignment algorithm calculates the attitude angles with the open-loop scheme and estimates the accelerometer and gyro biases with the closed-loop scheme at once. The estimated bias errors are used to correct the sensor errors that are utilized to calculate the attitude angles in the open-loop scheme. The computer simulation results illustrate the efficiency of this new alignment algorithm.