Research And Implementation Of Low-Cost Temperature Compensation Algorithm For MEMS Accelerometer

MEMS accelerometer is a key component in micro inertial navigation system,which plays an important role in the fields of aviation,automobile,intelligent sensing and control.It has the advantages of small volume,high precision,high integration and wide application.Due to the manufacturing process and material characteristics,MEMS accelerometer will always be disturbed by a variety of environmental factors in application,resulting in the performance degradation.Among them,the temperature has the greatest impact on the results.Importantly,the zero bias and scale factor are the most sensitive parameters.Therefore,reducing or eliminating the error caused by temperature is helpful for the application and cost down.This paper mainly focuses on the above problem by both the theoretical and experimental study.The main contents are as follows:Based on the working principle and temperature characteristics of MEMS accelerometer,three numerical algorithms including least square method,polynomial fitting method and neural network method are used for the analysis and research of temperature compensation.At the same time,their respective advantages and disadvantages are also compared.Then,combined with the experimental data,four numerical interpolation methods including piecewise linear interpolation,nearest neighbor interpolation,cubic polynomial interpolation and piecewise cubic Hermite interpolation,are studied and tested by judging the running speed and the accuracy.Finally,an improved polynomial fitting temperature compensation algorithm is proposed by combining the interpolation method with the traditional compensation algorithm.The test circuit of accelerometer temperature compensation is also built with the prepared software.After completing the design and writing of each interface timing protocol and Verilog code,the accelerometer can collect and compensate the acceleration data at different temperatures through the FPGA development board,and output the accelerometer compensation data results in the form of serial port in the upper computer.The interpolation improved polynomial fitting temperature compensation algorithm proposed in this paper is beneficial to increase more accelerometer data at various temperature.It both can reduce the cost by reducing the testing time,and improve the fitting accuracy by adding more data.The zero bias and scale factor of the accelerometer are compensated by this method,and then the final result of the accelerometer is compensated according to some parameters obtained by the compensation.The results show that the influence of temperature on the zero bias and scale factor is significantly reduced under the temperature range of-40℃～60℃ by the proposed compensation algorithm.Specifically,the root mean square error and range of zero bias are reduced by95.32% and 93.49% respectively;the root mean square error and range of scale factor are reduced by 98.75% and 89.29% respectively;the zero bias stability is improved by 95.2%and the temperature coefficient of scale factor is reduced by 81.3%.The fluctuation range of the actual acceleration value output by the final accelerometer is changed from 0.038 g to 0.003 g,which is reduced by 92.11%.The results show that this method can not only reduce the measurement cost,but also obtain high-precision compensation results,which is suitable for engineering application.