Calibration Analysis of High-G MEMS Accelerometer Sensor Based on Wavelet and Wavelet Packet Denoising

Area of Science:

• Signal Processing
• Instrumentation
• Mechanical Engineering

Background:

• High-G accelerometers are crucial for measuring motion in demanding fields like aerospace and projectile penetration.
• Piezoresistive accelerometers require effective denoising techniques to ensure accurate measurements.
• Wavelet analysis offers potential solutions for noise reduction in accelerometer data.

Purpose of the Study:

• To investigate and compare the efficacy of wavelet threshold denoising and wavelet packet threshold denoising for high-G piezoresistive accelerometers.
• To determine the optimal decomposition strategies for each denoising method.
• To evaluate the suitability of these methods for different accelerometer data processing tasks.

Main Methods:

• Adaptive decomposition and Shannon entropy criterion were employed to identify optimal decomposition layers and trees.
• Both wavelet threshold and wavelet packet threshold denoising utilized Stein unbiased likelihood estimation for soft thresholding.
• Numerical simulations and Machete hammer tests were conducted for validation.

Main Results:

• Wavelet threshold denoising demonstrated superior performance for the dynamic calibration of high-G accelerometers.
• Wavelet packet threshold denoising proved more effective for extracting parameters from the oscillation phase of accelerometer signals.
• The chosen adaptive decomposition and entropy criteria successfully optimized the wavelet analysis.

Conclusions:

• Wavelet threshold denoising is recommended for dynamic calibration applications of high-G accelerometers.
• Wavelet packet threshold denoising is the preferred method for detailed parameter extraction in oscillatory phases.
• The study provides valuable insights into selecting appropriate wavelet denoising techniques for specific high-G accelerometer data analysis needs.