Research On The Technology Of Calibration And Error Compensation For MIMU

MIMU(Micro Inertial Measurement Unit) is consisted of silicon micro mechanical gyroscope and micro accelerometer, it is the key component of micro inertial navigation system, and the accuracy of micro inertial navigation system directly depends on the measurement accuracy of MIMU. Compared with traditional inertial device, micro inertial device has lots of advantages such as smaller volume, smaller weight, lower power consumption, higher reliability, lower cost and so on, therefore, it presents a big applied foreground in the field of civil and military. However, as a small structure and the material of device is sensitive with temperature, micro mechanical gyroscope and micro accelerometer are easily affected by work environment and temperature, so a error of measurement is unavoidable. In order to increase the accuracy of micro inertial navigation system, a method must be fond to decrease system error, and a mass of studies show that more than 90% of the micro inertial navigation system error is caused by the error of MIMU. Build an error model to compensate system error is a effective method for increasing the accuracy of MIMU, but calibration is the precondition of error compensate, this article does some studies on the calibration and error compensate of MIMU.Firstly, analyzing the working principle and main performance parameter of MIMU and all kinds of error source, classifying the main error of MIMU into three groups, static error, temperature error and random error. Secondly, for static error, build error model for micro mechanical gyroscope and micro accelerometer, and calibrate the factor of main error with the method of “positive and negative rotation at six position”. Aim at the problem of MIMU is sensitive to temperature, analyzing the temperature character of micro mechanical gyroscope and micro accelerometer, comparing the advantages and disadvantages also the scope of application of several common temperature models, building a temperature model with reasonable mathematical model, through a temperature experiment, identifying the factor of temperature model for compensation with the method of multivariable nolinear regression. Then bring up a new method of calibration for whole MIMU, building a whole error model for micro mechanical gyroscope and micro accelerometer, concrete arrange and calculative formulas of calibration are given. Finally, about the random error, a random error model of micro inertial device is established by using stochastic signal processing methods and time series analysis theory in order to filter random noise by Kalman filter.