Evaluation of attitude estimation algorithms using absolute magnetic reference data: Methodology and results

Location is a growing problem due, for instance, to the expansion of civil drones or intelligent vehicles. This expansion was made possible through the development of microelectromecanical systems (MEMS), cheap and miniaturized inertial sensors. In this context, this article is devoted to the development of a simulator which generates the sensor measurements, giving a specific trajectory, in order to validate and compare pose estimation algorithms. After validation of the simulator with real movements and measurements, four pose estimation algorithms from the literature are compared on different trajectories. All these algorithms use magnetic field sensors in addition of the classical inertial sensors. This comparison is led to select an algorithm for a future application with multiple IMUs.