Filtering Algorithm of MEMS Gyroscope Based on Swing Markov Interacting Multiple Models

MEMS gyroscope is increasingly becoming important due to its advantages such as small, inexpensive, low power and reliable. However the accuracy of MEMS gyroscope is insufficient for many attitude determination applications as the large inherent noise. So many filtering algorithms are applied into the attitude compensations for MEMS gyroscopes, which are based on the assumption that the motions of attitude are relatively slow in a sampling interval. Unfortunately the assumption is unacceptable in the case of attitude motion with high frequency oscillations. This paper introduced Swing Markov model into describing the attitude oscillations and proposed Swing Markov Interacting Multiple Models (IMM) to establish the combinational filter for MEMS gyroscope. As Swing Markov model is correspond with the attitude oscillations, the results of simulation indicated the efficiency and accuracy of the proposed algorithm.