Performance Analysis of Electromyogram Signal Compression Sampling in a Wireless Body Area Network.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wireless Sensor Networks

Background:

• Increasing demand for portable hardware strains signal processing resources.
• Compressed sensing (CS) offers a promising solution for efficient signal acquisition in wireless sensor networks.
• High sample rate electromyogram (EMG) signals require optimized acquisition methods for reduced energy consumption.

Purpose of the Study:

• To comprehensively analyze the compressed sensing (CS) method for electromyogram (EMG) signal acquisition.
• To evaluate the performance of various sparse bases and reconstruction algorithms for compressed EMG signals.
• To identify optimal CS parameters for efficient EMG data processing and reduced energy usage.

Main Methods:

• A comparative analysis of 52 wavelet sparse bases and five reconstruction algorithms was performed.
• Performance was evaluated at different compression levels using metrics like percentage root mean square distortion (PRD).
• The study focused on sparse representation and reconstruction of high sample rate EMG signals.

Main Results:

• The db2 wavelet basis demonstrated effective sparsification of EMG signals, leading to proper reconstruction.
• The basis pursuit (BP) algorithm showed superior reconstruction efficiency and performance.
• Low PRD values were observed for the db2 wavelet basis across various compression ratios.

Conclusions:

• The db2 wavelet basis and BP reconstruction algorithm are suitable for sparse EMG signal processing.
• These findings provide a benchmark for practical CS applications in EMG signal acquisition.
• Optimized CS parameters can significantly reduce energy consumption and extend acquisition duration for EMG sensors.