Intelligent real-time MEMS sensor fusion and calibration

摘要

This paper discusses an innovative adaptive heterogeneous fusion algorithmbased on estimation of the mean square error of all variables used in real timeprocessing. The algorithm is designed for a fusion between derivative andabsolute sensors and is explained by the fusion of the 3-axial gyroscope,3-axial accelerometer and 3-axial magnetometer into attitude and headingestimation. Our algorithm has similar error performance in the steady state butmuch faster dynamic response compared to the fixed-gain fusion algorithm. Incomparison with the extended Kalman filter the proposed algorithm convergesfaster and takes less computational time. On the other hand, Kalman filter hassmaller mean square output error in a steady state but becomes unstable if theestimated state changes too rapidly. Additionally, the noisy fusion deviationcan be used in the process of calibration. The paper proposes and explains areal-time calibration method based on machine learning working in the onlinemode during run-time. This allows compensation of sensor thermal drift right inthe sensors working environment without need of re-calibration in thelaboratory.