Analysis, Modeling and Compensation of Bias Drift in MEMS Inertial Sensors

Inertial sensors have a broad field of applications and are especially essential for localization. One of the most severe errors of MEMS inertial sensors is bias drift. In this paper bias drift is mathematically modeled as a combination of time and temperature dependent behaviors which are fused together allowing online bias compensation. The example presented indicates an almost perfect estimation resulting in elimination of more than 99% of bias. In this paper, also, a novel technique for modeling systems with hysteresis is presented.