A Coarse Alignment Method for Gravity Vector Reconstruction Based on Adaptive Kalman Filtering

In this paper, a coarse alignment method using gravity vector reconstruction based on adaptive Kalman filtering for strapdown inertial navigation system (SINS) is proposed. Firstly, the change relationship between the real gravity vector and time is derived. The gravity vector after noise reduction can be constructed to improve the accuracy of the gravity vector. Next, the gravity vector is reconstructed based on the adaptive Kalman filtering to obtain a new vector observer model. Then, the improved quaternion Kalman filtering algorithm is proposed to reduce the influence of pseudo-measurement noise, the initial coarse alignment process is realized by using the quaternion chain rule. Finally, the simulation experiments are designed to verify the performance of the proposed method. The alignments results show the proposed method can filter out the random noise well and has better smoothness.