Relationship of the accuracy of estimation of the constant error component of the MEMS accelerometers with the Allan variation

The article has established the relationship between the Allan variation and the estimation error variance of the constant component of output signal of the inertial sensor obtained at averaging. It is shown that under certain conditions the Allan variation coincides with the estimation error variance of a constant component. This allows through the use of the Allan variation curve, insensitive to a constant component to estimate the minimum value of the estimation error variance of a constant component by the method of averages and to determine averaging optimal time which is important when calibrating inertial sensors. Thus, according to the Allan variation curve, which is insensitive to the presence of a constant component, it is possible to draw conclusions as to the accuracy of its evaluation. In this case optimal averaging time as to the criterion of minimum variance is directly proportional to the square root of the ratio of intensity of white noise to intensity generating noise of the Wiener process and the error variance is directly proportional to their multiplication and square of the instability factor of the output signals zero shifting. According to this method, which is based on using the Allan variation, a specialized software has been developed and tested that allows computing the Allan variation on the results of processing of high-frequency data inertial sensors (accelerometers and gyros) in the long lasting launches from the subsequent polynomial approximation. An assessment of noise components of the output signals of MEMS accelerometers Gemini CAS211/291 in three subsequent launches followed by averaging time appraisal for each accelerometer axis and calculation of minimum variance and mean square deviation of the estimation errors of constant components at their calculation by averaging have been accomplished. Later in the research of accuracy characteristics of navigational instruments and appliances at instrument making plants it was proposed to use both the classical methods of statistical analysis and the Allan variations method as an effective way to identify the noise patterns in the channel of a measurement means