GPS/SINS Speed Integral Matching Alignment Based on Improved Adaptive UKF

摘要

To solve the problem of system error and environmental interference of MEMS inertial sensor, which make SINS cannot quickly and accurately alignment, a GPS/SINS speed 'integral' matching alignment method is designed. Firstly, the GPS velocity measurement values are used to assits the SINS for coarse alignment in this paper. Then, speed integral matching fine alignment is performed. Establishing the alignment model, the solution speed of the SINS is corrected by the least squares curve fitting method and the arm effect is compensated for the GPS speed signal before filtering. Using the unscented Kalman filter (UKF) and the improved adaptive UKF to achieve the fine alignment respectively. Outdoor experiments and simulation's results show that the improved adaptive UKF can achieve higher alignment accuracy and speed. Compared with the common UKF, the accuracy of the east, north and up misalignment angles is inhanced by 0.12°, 0.03° and 0.19°, respectively. The alignment time is shortened by 20s, 10s and 25s, respectively. The above work can provide a reference for AUV accurate navigation.