Consistent calibration of magnetometers for nonlinear attitude determination

摘要

Three-axis magnetometers (TAMs) have been widely utilized as the cornerstone of integrated navigation (IN) and attitude determination (AD) in many aerospace systems. However, accurate navigation and AD demands for precise calibration of TAM. For this purpose, a complete TAM calibration process is presented in the current research to compensate all of the key errors. In this regard, a hyper least square (HyperLS) estimator is extended for accurate and consistent ellipsoid fitting problem of TAM calibration. Subsequently, the calibrated TAM is utilized for real time attitude determination via non-linear colored noise filters of extended Kalman filter, simplex unscented Kalman filter (SUKF) and cubature Kalman filter. Effectiveness of the proposed calibration scheme is investigated through implementation on a Honeywell HMC5883L magnetometer. Comparative study of the results with those of well-known methods of strict ellipsoid and the three-step algorithm demonstrates the suitability of the HyperLS and SUKF in terms of accessible estimation accuracy. (C) 2015 Elsevier Ltd. All rights reserved.