Joint Calibration of Inertial Sensors and Magnetometers using von Mises-Fisher Filtering and Expectation Maximization

Microelectromechanical-systems-based inertial sensors and magnetometers are low-cost, off-the-shelf sensors that are widely used in both consumer and industrial applications. However, these sensors suffer from biases and effects such as axis misalignment or scale errors, which require careful system design and periodic sensor calibration. In this paper, we propose a fast calibration method for jointly calibrating inertial sensors and magnetometers based on discrete-time von Mises-Fisher filtering and expectation maximization. We evaluate the method on built-in smartphone sensors and show that the proposed method efficiently estimates the sensors' parameters and reduces the overall error.