Random Noise Compensation Method for MEMS Accelerometer Based on EMD Decomposition

In order to improve the measurement accuracy of MEMS accelerometers, a stochastic error compensation method based on empirical mode decomposition (EMD) is adopted. In this paper the accelerometer signal is decomposed into intrinsic mode function (IMFs) and a residual component by EMD algorithm, and IMF is divided into three categories: noise dominant component, signal/noise mixed component and signal dominant component; noise reduction of signal/noise mixed component is realized by threshold processing, the signal/noise mixed component after noise reduction and the dominant component of the signal are reconstructed to obtain the accelerometer signal after noise reduction. It is proved that the algorithm can effectively improve the measuring accuracy of the accelerometer through the verification of dynamic data and static data acquisition.