A method of estimating the silicon MEMS Gyro's drift base on multi-scale and multi-parameter

Because there were a lot of factors affect the precision of MEMS-Gyro, so arming at the output of non-stationary nonlinear characteristics, it was difficult to build the error model of MEMS-Gyro. An improving has been made on the wavelet neural network method. A method of multi-scale makes the output stationary and multi-parameter nonlinear estimated the drift of Gyro. The experimental results show that the precision of the MEMS-Gyro has been improved from 1°/s to 0.02°/s by the method of multi-scale and multi-parameter.