Abstract:

An improved adaptive square root filtering algorithm for solving transfer alignment under time varying noise is designed. In the algorithm, the state covariance adjusting factor matrix, adaptive adjusting system noise and measurement noise covariance, as well as the infinite noise memory scale are incorporated into the square root filtering structure which is computed by a sequential filtering algorithm. Through onestep controlling and multi step adaptive adjusting process, the performance in numerical computing, noise refraining and adaptive adjusting of the algorithm is remarkable improved with low calculation cost. Simulation results show that the algorithm possesses strong stability, high filtering accuracy and is capable of quickly adjusting according to the real noise, which provides an effective rapid and accurate transfer alignment method.