A magnetometer error compensation method based on external attitude assistance

Abstract The magnetic field serves as a crucial source of navigational information, typically measured by magnetometers in practical applications. However, magnetometers are inherently unstable instruments that require error compensation before formal use. Drawing inspiration from the system-level calibration approach of gyroscopes, this paper proposes a system-level error compensation method for magnetometers. By leveraging high-precision external attitude information, this method effectively measures the installation errors between the magnetometer’s three axes and the carrier coordinate system. It simultaneously calibrates sensitivity errors, non-orthogonality errors, and bias errors. The problem is formulated as a state estimation task and solved using an extended Kalman filter. Based on the estimation results, the compensation of the magnetometer was completed. The improvement of the magnetometer modulus and heading angle before and after compensation demonstrated the superior performance of the proposed method.