Research On Error Compensation Technology Of MEMS Acceierometer

In recent years,MEMS accelerometer has been widely used in our daily life because of its small volume,light weight,low power consumption and strong reliability.However,MEMS accelerometer has the disadvantage of low output data accuracy.Therefore,improving the accuracy of MEMS accelerometer has very important realistic meaning.For low accuracy problems of MEMS accelerometer,this paper designs and builds a complete set of calibration system,and focuses on solving the deterministic error and random error of MEMS accelerometer.The main contents of the work completed are as follows:(1)The overall design of MEMS accelerometer calibration system is carried out,and the inertial unit adis16488 a and angle sensor are selected.(2)Based on C# and WPF language,the host computer software design of the calibration system is completed,the 3D matrix transformation and wpf3 d related knowledge are studied,and the accurate modeling of the three-axis turntable is completed through 3DMAX.The precise control of the three-dimensional rotation of the turntable model can be realized through the operation of the front-end interface of the software.(3)Complete the hardware circuit design of the lower computer part of the calibration system,and carry out the welding and testing of PCB plate making and related electronic components.(4)Keil5 is used to complete the design and implementation of related drivers of the calibration system,including SD card drive,inertial unit drive and so on.Realize the automatic configuration of Wi Fi communication parameters in Wi Fi module,the calibration and automatic calibration process control of three-axis turntable,and complete the overall debugging of the PC software and lower computer of the Calibration system.(5)The static six position method is used to eliminate the definite error.The definite error source of MEMS accelerometer is analyzed and the error model is established.The collected output data is used to solve the error coefficient of the model,and then replaced with the error compensation model.The output data accuracy of the 18 positions after compensation is at least0.03 g higher than that before compensation.(6)Compensation of random errors using time series analysis,using Kalman filter for error noise reduction,and the Allan variance method is used to analyze the error noise source of MEMS accelerometer.The results before and after filtering show that the filter based on time series model can effectively reduce the random noise of MEMS accelerometer,The acceleration random walk is reduced by 92% and the speed random walk is reduced by10% compared with that before filtering.