A smart approach to EMG envelope extraction and powerful denoising for human-machine interfaces

Area of Science:

• Biomedical Engineering
• Signal Processing
• Human-Machine Interfaces

Background:

• Electromyography (EMG) is crucial for human-machine interfaces (HMIs), measuring muscle contractions via EMG envelopes.
• Raw EMG signals are susceptible to powerline interference and motion artifacts, compromising HMI reliability and performance.
• Advanced filtering techniques offer high performance but are often unsuitable for resource-constrained platforms.

Purpose of the Study:

• To investigate the efficacy of feed-forward comb (FFC) filters in denoising raw EMG signals.
• To assess the FFC filter's capability in removing both powerline interference and motion artifacts.
• To evaluate the suitability of FFC filters for low-cost, low-power HMI applications.

Main Methods:

• Application of feed-forward comb (FFC) filters to raw EMG signals.
• Implementation of FFC filter and EMG envelope extraction without multiplication operations.
• Offline performance evaluation using simulated noisy EMG signals and real-world noisy EMG data.
• Real-time testing on an Arduino Uno board.

Main Results:

• FFC filters effectively removed powerline interference and motion artifacts from EMG signals.
• Correlation coefficients for filtered EMG envelopes exceeded 0.98 (powerline noise) and 0.94 (motion artifacts).
• Successful real-time implementation and operation demonstrated on a basic microcontroller.

Conclusions:

• FFC filters provide a computationally efficient and effective solution for denoising EMG signals.
• This approach significantly improves the reliability of EMG-based HMIs, especially on low-power platforms.
• The proposed method offers a viable alternative to complex filtering for practical HMI applications.