Modeling Techniques And Stability Testing For Accelerometer Zero Offset

As an important component for sensing the carrier acceleration in inertial motion vector system, zero offset stability of quartz flexible accelerometer(hereinafter referred to as: accelerometer) in the storage period is an important indicator to test the performance, because the length of zero offset stability will indirectly impact the efficacy and maintainability of the inertial motion vector system. Therefore, it plays an important significance for researching the zero offset stability under the storage conditions that designing accelerometer data acquisition device and modeling an right mathematical model under a storage condition.A multi-sensor data acquisition system, which is on the base of Zig Bee protocol wireless networking, is designed. The network data acquisition system is used for collecting the data of accelerometer in eight locations in the storage period, and the least-square intercalation and wavelet noise reduction method is used to testing respectively attitude symmetry stability, zero offset repeat stability and zero offset drift period of zero offset of accelerometer.In order to research the long-term stability of accelerometer under the storage condition, the EMD time series model is selected to be mathematical model for long-term researching stability. Firstly, quartz flexible accelerometer in the storage conditions for zero offset data time series modeling, the modeling results, the model prediction accuracy is not high on the relevant zero offset data, the prediction accuracy increases with the shortcomings of error prediction time is bigger, so the development of a method based on empirical mode decomposition(referred to as: EMD) improved combined with the model of time series forecasting model, through the EMD algorithm with length dependent bias data was decomposed into several intrinsic mode with short related components, and then for each intrinsic mode component model, to solve the traditional time series prediction accuracy of long-range zero offset data, improve the prediction accuracy of the long time. Finally, through time series model simulation of the traditional time series model and improved in this paper, the simulation results show that: the EMD model and the ARMA model developed in this paper based on the study of quartz flexible accelerometer zero offset stability data is valid.