Design of On Board Calibration Algorithms of Satellite Magnetometer based on Two Stage Centered Solution and Kalman Filter Methods

Magnetometer is one of the most important sensors used in the satellite attitude determination and control system. Due to occurrence of various errors when the satellite is separated from the launcher and also during its rotation in the orbit, it is necessary to re-adjust onboard the sensor parameters. For this purpose, some solutions are proposed in this paper in which the satellite current attitude is not required. In this regard, first a magnetometer model is presented that despite conventional models; it includes nonlinearity, hysteresis and data quantization effects, permeability and installation error. Then, for sensor onboard calibration purposes, two stages-offline and two-stage online series structures are suggested. In the offline case, the centered solution and Levenberg Marquardt methods have been integrated. Also, the extended and unscented Kalman filters are integrated for online case. Utilizing the suggested algorithms, different errors including bias, scale factor and installation errors are simultaneously determined and also the accuracy is improved compared to the similar works. The simulation results for a Leo satellite show that the sensor parameters are derived with acceptable accuracy. Accordingly, it will be illustrated that the centered solution method has lower computational load and shorter time convergence, but it has lower accuracy with respect to online methodology.