An equivalent backtracking coarse alignment method with dynamic optimal sliding window integration

Coarse alignment is a critical step in the initial alignment of a strapdown inertial navigation system. We propose an equivalent backtracking coarse alignment method with dynamic optimal sliding window integration to address the limited accuracy of existing coarse alignment methods and the storage constraints of backtracking coarse alignment techniques. Firstly, the paper presents the equivalent backtracking coarse alignment method, derives the update method for the constant attitude matrix, and demonstrates the equivalence between traditional and equivalent backtracking coarse alignment algorithms. Simultaneously executing forward and backtracking alignment calculations reduces the navigation system's storage requirements and overall cost. Secondly, a dynamic optimal sliding window integration strategy quantifies the observation vector error corresponding to various sliding window lengths. This optimization improves the observation accuracy and the algorithm's robustness, enhancing alignment precision. Finally, high-precision compensation for initial velocity is applied to enhance the accuracy of the observation vector, further improving alignment precision. Semi-physical simulation experiments validate the effectiveness and correctness of the proposed algorithm.