High-G MEMS Accelerometer Calibration Denoising Method Based on EMD and Time-Frequency Peak Filtering.

Area of Science:

• Instrumentation and Measurement
• Signal Processing

Background:

• Accelerometer calibration is susceptible to noise, impacting data accuracy.
• Existing denoising methods may compromise signal characteristics.

Purpose of the Study:

• To propose and validate a new accelerometer denoising method.
• To effectively suppress noise during accelerometer calibration.
• To preserve original signal characteristics post-denoising.

Main Methods:

• Empirical Mode Decomposition (EMD) to decompose the signal into intrinsic mode functions (IMFs).
• Time-Frequency Peak Filtering (TFPF) applied to medium-frequency IMFs.
• Selective removal of high-frequency IMFs and preservation of low-frequency IMFs.
• Signal reconstruction and Allan variance analysis for performance evaluation.

Main Results:

• The proposed EMD + TFPF method significantly suppresses random noise during calibration.
• Signal reconstruction error is controlled within 0.5%.
• The filtering effect is substantial, with a 97.4% improvement over original data.

Conclusions:

• The EMD + TFPF algorithm offers an effective solution for accelerometer denoising.
• This method successfully removes noise while preserving essential signal features.
• The approach demonstrates superior performance compared to other methods evaluated by Allan variance.