Methodology Proposal of EMG Hand Movement Classification Based on Cross Recurrence Plots

Area of Science:

• Biomedical Engineering
• Signal Processing
• Nonlinear Dynamics

Background:

• Electromyography (EMG) signals present challenges due to their nonstationary, noisy, and high-dimensional nature, complicating accurate prediction and classification.
• Cross recurrence plots (CRPs) are effective in identifying subtle patterns within noisy signal environments.
• Traditional machine learning methods may struggle with the inherent complexities of EMG data.

Purpose of the Study:

• To propose and validate a novel methodology for detecting and classifying surface EMG (sEMG) signals from hand movements.
• To assess the accuracy, sensitivity, and specificity of the proposed method in avoiding false classifications.
• To compare the performance of the novel method against traditional machine learning approaches.

Main Methods:

• Fifty subjects performed ten distinct hand movements, with EMG data collected from electrodes placed on each arm.
• Nonlinear features of the sEMG signals were analyzed using cross recurrence quantification analysis (CRQA).
• A new methodology was developed, centering on CRQA for movement detection and classification, augmented with tools to ensure classification robustness.

Main Results:

• The proposed CRQA-based methodology achieved good accuracy, sensitivity, and specificity in classifying sEMG signals.
• The method effectively detected and classified ten different hand movements across fifty subjects.
• CRQA demonstrated advantages over traditional machine learning methods in classifying complex EMG patterns.

Conclusions:

• The developed CRQA methodology is a feasible and effective technique for classifying sEMG signals.
• This approach offers improved performance and robustness compared to conventional machine learning methods for EMG analysis.
• The findings highlight the potential of nonlinear dynamics, specifically CRQA, in advancing EMG-based human-computer interfaces and movement analysis.