A New Compressed-Structure-Based Coning Algorithm for Fiber Optic Strapdown Inertial Navigation Systems

Abstract

For strapdown inertial navigation systems (SINSs) equipped with high-precision fiber optic gyroscope whose outputs are angular rate signals, the attitude calculation accuracies of traditional angular-rate-based rotation vector algorithms are limited when the vehicle is in a highly dynamic environment due to the fact that only the compensation of the second-order noncommutativity error is considered. Although the attitude algorithms based on polynomial iteration have higher accuracy, the computational complexity is too large. In order to fully tap the accuracy of the inertial measurement unit (IMU) to reduce the attitude errors, the third-, fourth-, and fifth-order Picard solutions for the equivalent rotation vector differential equation are derived in this article. A new method with a compressed structure to calculate the rotation vector for real-time processing systems is proposed to further reduce the noncommutativity error. In a classical coning motion environment, the frequency-series-based Taylor expansions are utilized for error analysis and design of the new algorithm’s coefficients. The effectiveness of the proposed coning algorithm with eighth-order accuracy is validated by digital simulation and initial alignment experiments in the turntable. In comparison with the traditional algorithms, the proposed algorithm improves the yaw alignment accuracy by 78.5%, the pitch alignment accuracy by 73.8%, and the roll alignment accuracy by 66.9% on average.