چكيده به لاتين
The existence of high-level stochastic errors in the output of microelectromechanical system (MEMS)-based gyroscopes has restricted their use in high-precision applications, including inertial navigation. The virtual gyroscope technology improves the accuracy of the MEMS-based gyroscopes by combining the multiple outputs of the gyroscope array. In this thesis, due to the weakness of the virtual gyroscope technology in improving the measurement accuracy in dynamic conditions, improving the dynamic characterization of the system is considered as a tool for obtaining higher measurement accuracy and a new dynamic model for modeling the bias drift of a MEMS-based gyroscope is obtained. In contrast to the conventional dynamic model for describing the behavior of the bias drift of the MEMS-based gyroscopes, the proposed model, gives an observable description for the system. Then, by extending the proposed model for a gyroscope array consisting of four MEMS-based gyroscopes, the measurements are optimally combined using the Kalman filter (KF) to improve the measurement accuracy of the virtual gyroscope. For a practical KF implementation, obtaining the correlation coefficients between the stochastic errors of multiple gyroscopes in a gyroscope array is very important. For this purpose, a mathematical statistics method is proposed for obtaining the practical correlation factors between the stochastic errors. The proposed method is also able to determine the practical values of the process and measurement noise covariance matrices for a practical KF implementation. By obtaining the practical correlation coefficients, it was also possible to evaluate the correlation effect on the accuracy improvement of the virtual gyroscope measurement. Theoretical analyses and simulations have shown that the proposed dynamic model is more effective than the conventional dynamic model in improving the measurement accuracy of the virtual gyroscope under dynamic conditions.
Keywords: MEMS-based gyroscope, virtual gyroscope technology, data fusion, wavelet decomposition, correlation identification
