Error Compensation of Three-Axis Magnetic Sensor Based on Improved Ellipsoidal Fitting Method

Aiming at the errors generated by the three-axis magnetic sensor during measurement, this article proposes a new algorithm: an ellipsoid fitting algorithm with an “outlier removal and random iterative selection” mechanism to compensate the magnetic sensor. To address the problem that traditional compensation algorithms cause deviations in the fitting results due to the consideration of data with significant errors, an improvement is implemented by adopting the concept of random selection. A subset of the data is randomly selected from the original measurement data, and these data are fitted into an ellipsoid. The distance of each point to this ellipsoid surface is calculated to determine the new matching points. This process is cycled multiple times to select the optimal fitting ellipsoid. This approach can accurately eliminate data with large errors and effectively enhance the error compensation effect of the sensor.