A novel magnetometer-accelerometer calibration based on a least squares approach

In this work a calibration method which is easy to use and is capable of calibrating a combination of a magnetometer and an accelerometer is described. The calibration method accounts for hard and soft iron effects created by the platform, sensor errors like biases, scale factors and non-orthogonalities as well as the relative orientation between the magnetometer and the accelerometer. The algorithm is based on a least squares problem, which is solved using a singular value decomposition. During the calibration process the spatial orientation of the platform is not needed. Therefore no additional hardware is required. The functionality and accuracy of the algorithm is shown using simulated sensor readings.