Noise Reduction for MEMS Gyroscope Signal: A Novel Method Combining ACMP with Adaptive Multiscale SG Filter Based on...

Area of Science:

• Signal Processing
• Sensor Technology
• Mechanical Engineering

Background:

• Micro-electromechanical system (MEMS) gyroscopes are crucial for navigation and control.
• Signal noise significantly degrades MEMS gyroscope performance.
• Existing denoising methods often struggle with complex noise patterns in MEMS gyroscope signals.

Purpose of the Study:

• To propose a novel hybrid denoising method for offline MEMS gyroscope signals.
• To improve the accuracy and reliability of MEMS gyroscope data.
• To offer a superior alternative to current signal denoising techniques.

Main Methods:

• A hybrid method combining adaptive chirp mode pursuit (ACMP) and adaptive multiscale Savitzky-Golay filter (AMSGF) is introduced.
• ACMP with a modified mutual information-based stopping criterion decomposes the signal into modes.
• AMSGF, utilizing adaptive moving average (AMA) for adaptive window sizing, further processes these modes.

Main Results:

• The proposed ACMP-AMSGF method demonstrates superior denoising performance compared to EMD, discrete wavelet, and VMD methods.
• Practical denoising results under various motion conditions validate the method's effectiveness.
• AMSGF shows better denoising capabilities than other common smoothing techniques.

Conclusions:

• The novel hybrid ACMP-AMSGF method offers an effective solution for offline MEMS gyroscope signal denoising.
• The adaptive nature of the proposed filters enhances noise reduction while preserving signal integrity.
• This method provides a significant advancement in MEMS gyroscope signal processing.