Review on electromyography based intention for upper limb control using pattern recognition for human-machine...

Area of Science:

• Rehabilitation Engineering
• Biomedical Signal Processing
• Human-Machine Interface

Background:

• Upper limb myoelectric prosthetic control is crucial for restoring function.
• Electromyogram (EMG) signals, including surface (sEMG) and intramuscular (iEMG), are key for prosthesis control.
• EMG analysis is vital for machine learning, deep learning, and understanding neural mechanisms.

Purpose of the Study:

• To review current techniques in EMG signal processing for prosthetic control.
• To discuss electrode array utilization, signal acquisition, preprocessing, feature extraction, and classification.
• To explore alternative muscle activity measurement methods and future deep learning applications.

Main Methods:

• Review of existing literature on EMG signal processing and classification.
• Analysis of pattern recognition techniques like Support Vector Machine (SVM), k-Nearest Neighbor (KNN), and Bayesian classifiers.
• Discussion of electrode array configurations, signal conditioning, feature engineering, and dimensionality reduction.

Main Results:

• Established pattern recognition methods effectively classify EMG signals for prosthetic control.
• Various techniques for signal acquisition, preprocessing, and feature extraction are employed.
• Alternatives like force sensors offer reliable muscle activity measurement.

Conclusions:

• EMG signal processing is a mature field with established classification methods.
• Future advancements in deep learning, such as Artificial Neural Networks (ANN) and Recurrent Neural Networks (RNN), promise more robust EMG classification.
• Continued research in EMG signal analysis will enhance prosthetic functionality and human-machine interaction.