Model design based on MEMS gyroscope random error

A model is made in view of the MEMS gyroscope random error, which is applied to error compensation with the Kalman filter. And main noise sources that affect measurement accuracy are determined via Allan variance method. The correctness of the model is verified by data filtering, proper error model and error compensation of the MEMS gyroscope. The principle factors that affect the performance of MEMS gyroscope are confirmed with the analyses of MEMS gyroscope noise items and the coefficients of various noise sources are compared before and after filtering using Allan variance method, the experiment shows that error model significantly improved the precision of the measurement of MEMS gyroscope.