The Research Of MEMS Inertial Devices Error Analysis And Compensation Method

MEMS inertial devices are widely used in navigation and guidance system with small size, low power consumption and low cost. But because of the low precision, poor performance compared to other inertial device cause it has limitation in practical engineering application. In this paper, to improve the low cost MEMS inertial devices through actual use precision test and error compensation based on the existing condition, has the extremely vital significance.First of all, analyze the MEMS inertial device system error. According to the current equipment in laboratory to determine the static method and dynamic method respectively for calibration of gyroscope and accelerometer, build the error model of the system and do compensation, to verify the accuracy of error model of normal temperature environment.Secondly, do the full temperature test on MIMU, according to the inertial devices error coefficient under different test condition is the strong non-linear characteristic,design the RBF neural network. To train the device zero bias, scale factor and so on, to improve the environment condition of the system to use precision.Thirdly, using Allan variance method to identification and characterization the random error. The results of noise source analysis made foundation for the random error of the filter to improve the precision of system error.Finally, using the Kalman random filtering, inspection and processing of random data. Then set up ARMA model, through the linear Kalman, adaptive Kalman filtering for random signal filtering. By testing the compensation, the results in the all conditions achieve the precision requirement.