Experimental Research on feedback Kalman Model of MEMS Gyroscope

The time sequence analysis method, such as ARMA model, was usually adopted in creating drift error model for MEMS gyroscope. This model decreases drift error a lot by making use of Kalman filter algorithm based on ARMA model, but, at the same time, its dynamic feature will be depressed. In this paper, we proposal a novel Kalman method which change Q through feedback. Though this method has the same effect on static state, it improve dynamic performance greatly. We has implemented it on FPGAs, and has did some experiment research on one gyroscope, the result show that this method can improve static and dynamic feature simultaneously.