This paper presents data-driven self-calibration control strategy based on variable expectation-maximum (VEM) algorithm to improve accuracy of MEMS gyroscope under unknown stochastic disturbances in harsh environment. Disturbances sourcing from device insides with slow-varying, long-duration characteristic and external environment with drastic-magnitude, short-duration disturbances characteristic are successively constructed in the two measurement models of the gyroscope, respectively. The controller of the presented method is dynamically chosen to identify these disturbances to reduce the bias variance of the gyroscope. Numerical calculation results based on experimental test shows that the bias variances are reduced to 48.654°/s, 3.981°/s, 0.587°/s by Kalman filter (KF), least mean square (LMS), standard EM (SEM) methods, respectively. Comparing with these algorithms, bias variance of the proposed method is improved by 287.9, 23.6 and 3.5 times, respectively. It is proved that the proposed method can improve the accuracy by two orders of magnitude.
Data-Driven Self-calibration Control Strategy for Accuracy Enhancement of Micromachined Gyroscope with Two Orders of Magnitude
Lect. Notes Electrical Eng.
2021-10-30
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Micromachined resonant gyroscope [5253-26]
British Library Conference Proceedings | 2003
|A New Micromachined Gyroscope with Digital Output
British Library Conference Proceedings | 2003
|A new micromachined gyroscope with digital output
Tema Archiv | 2003
|Micromachined Gyroscope in Silicon-On-Insulator (SOI) Technology
British Library Conference Proceedings | 2002
|