An Online Combined Compensation Method of Geomagnetic Measurement Error

Abstract

The online error compensation of the three-axis magnetic sensor is the key to a high accuracy geomagnetic assisted navigation. Traditionally, the least squares or its improved methods are used for error compensation, which are offline and need to assume that measurement errors are Gaussian. In this paper, the geomagnetic measurement error is modeled based on the error source analysis, and then the compensation parameters are converted through the ellipsoid hypothesis. To eliminate the negative effect of abnormal values, iterative least squares is designed to calculate static compensation parameters and corresponding residuals, then different weights are assigned to each geomagnetic measurement data through robust estimation. Also, to avoid data saturation phenomenon and finally complete online error compensation, a forgetting factor is introduced into the iterative least squares, and the dynamic coefficients are constructed by weights of measurement data to derive the online updated compensation parameters. Simulation and experiment results both show that the online combined compensation method possesses higher adaptability, which can eliminate the influence of abnormal data effectively. And its accuracy and stability of error compensation are better than other state-of-the-art methods.