A MEMS-based Magnetometer Calibration Approach in AUV Navigation System

Abstract

MEMS-IMU has the characteristics of small size and low cost. It can greatly reduce the cost when applied to underwater navigation and positioning system. However, MEMS-IMU has lower precision and poorer stability, especially MEMS magnetometers are more susceptible to the surrounding magnetic field environment. Therefore, it is of great significance to study a stable and effective calibration method for magnetometers. In this paper, by analyzing the existing magnetometer calibration methods, namely the ellipsoid fitting method and the dot product invariance method, an improved method namely multi-constraint calibration based on two is proposed. Based on the three basic properties of the geomagnetic field, the error equation of the multi-constraint calibration method is established, the nonlinear optimization objective function is obtained from the error equation, and the nonlinear target optimization function is solved by the Levenberg-Marquardt (L-M) method to obtain the calibration parameters. The error is compensated to obtain an ideal magnetometer output, which improves the accuracy of the AUV(Autonomous Underwater Vehicle) underwater navigation system. The simulation results show that the multi-constraint calibration method has obvious advantages compared with the ellipsoid fitting method and the dot product invariant method in solving the calibration parameters.