A rapid SINS alignment method based on observational measurement extension

Abstract Initial alignment is one of the critical technologies for the smooth and effective operation of the strapdown inertial navigation system (SINS), and the alignment time and accuracy are the two deterministic indicators of SINS initial alignment performance. To address the issues of redundant alignment time and low accuracy in the initial alignment process, a new method utilizing system observations is proposed, which includes velocity errors, equivalent specific force outputs, and equivalent gyro angular velocity errors. System state equations and observation equations are established based on the fundamental principles and error characteristics of inertial sensors. Singular value decomposition was employed to calculate the singular values of each system’s observation matrix, and observability was determined through singular value analysis. Experimental results indicate that the inclusion of equivalent specific force outputs accelerates the convergence of the horizontal misalignment angle, while equivalent gyro angular velocity errors expedite the convergence of the azimuth misalignment angle. Proposed method with enhanced system observability improves alignment accuracy significantly, with horizontal accuracy increased by approximately 62% and vertical accuracy by about 42% compared to a filtering model using only velocity errors as observations.