A Fast SINS Self-Alignment Method Under Geographic Latitude Uncertainty

Abstract

To achieve high alignment accuracy of strapdown inertial navigation system (SINS) with the limited time and the unknown geographic latitude constraints, a novel fast self-alignment method with the real-time estimated latitude is proposed. In the stage of the coarse alignment, the latitude self-estimation and the coarse alignment are performed at the same time. The geometric relationship between the transition angle and the latitude is established, where the angle can be determined by the gravity acceleration vector at different times. Based on the reconstructed observation vector and the set dynamic window, a real-time latitude self-estimation algorithm is proposed for a swaying case, and the attitude determination procedure is accomplished by the improved filter quaternion estimator (filter-QUEST) algorithm, where the vectors are constructed based on a generalized integration regulation. With the deduced backtracking alignment error model, the forward fine alignment is performed with the statistical latitude estimate and the saved data during the process of the coarse alignment, thus will obviously speed up the overall alignment time. The results of the mathematical simulation and the turntable experiment are performed to validate the estimation and alignment performance of the proposed approach.