Myoelectric Temporal Patching: Future Prosthetics Shall Effectively Leverage sEMG Temporal Patterns.

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Signal Processing

Background:

• Robust myoelectric control is crucial for lower limb assistive devices and prosthetics.
• Current deep learning methods for surface Electromyographic (sEMG) signal feature extraction have high computational demands, limiting their use in prosthetics.
• Effective feature extraction from sEMG is key to improving the performance of myoelectric controllers.

Purpose of the Study:

• To introduce a novel, computationally efficient handcrafted feature extraction method called Myoelectric Temporal Patching (MTP).
• To capture short- and long-term temporal dynamics in sEMG signals without significant computational overhead.
• To validate the effectiveness of MTP in recognizing lower-limb motion intentions for prosthetic applications.

Main Methods:

• Developed the Myoelectric Temporal Patching (MTP) method for extracting multi-signal features from sEMG signal segments.
• Applied MTP to extract temporal features, propagating information across windows to capture dynamics.
• Evaluated MTP using two pattern recognition experiments: gait phase recognition (SIAT-LLMD dataset) and locomotion mode recognition (MyPredict 1 dataset).
• Compared MTP performance against traditional and spatial features using Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), and Support Vector Machine (SVM) models.

Main Results:

• The MTP feature set significantly outperformed traditional and spatial features across all tested machine learning models (p < 0.0001).
• Achieved peak accuracies of 85.11% for gait phase recognition and 87.70% for locomotion mode recognition using SVM with MTP features.
• Demonstrated the ability of MTP to effectively decode lower-limb motion intentions by leveraging sEMG temporal dynamics.

Conclusions:

• The proposed Myoelectric Temporal Patching (MTP) method offers a robust and computationally feasible solution for sEMG feature extraction.
• MTP's effectiveness in capturing temporal dynamics is critical for advancing lower-limb prosthetic and assistive device control.
• This work supports the development of commercial-level lower limb assistive devices by providing efficient and accurate myoelectric control solutions.