Rapid alignment for SINS based on expansion of observation vectors in polar regions

Abstract

In polar regions, the rapid convergence of meridians and the small horizontal component of Earth's rotation make traditional North-pointed initial Alignment algorithm (NA) ineffective. While Alignment algorithm based on Pseudo-Earth frame (PA) can solve the problems caused by the rapid convergence of meridians, the observability of Strap-down Inertial Navigation System (SINS) on stationary base is so poor that it takes much time to complete initial alignment. According to this case, in this paper, a novel alignment algorithm based on expansion of observation vectors is proposed to implement rapid initial alignment in polar regions. This paper mainly completes the following works: firstly, a rapid Alignment algorithm based on Expansion of Observation vectors for Pseudo-earth frame (EOPA) is proposed, then the reason for its rapid convergence is explicated by analysis of the error equations; secondly, two observability analysis methods: Singular Value Decomposition (SVD) and the Attitude Dilution Of Precision (ADOP) are adopted to compare the observability of PA with that of EOPA. Finally, semi-physical simulation shows that in polar regions, even in the case of latitude of 89° and initial misalignment angle of 3°, the azimuth misalignment error can be reduced within 0.25° in 40s. It has significant engineering application value.