Estimation of Knee Extension Force Using Mechanomyography Signals Based on GRA and ICS-SVR.

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Signal Processing

Background:

• Noninvasive muscle force estimation is crucial for real-time adjustment of assistive forces during lower-extremity rehabilitation.
• Mechanomyography (MMG) signals offer a noninvasive alternative to surface electromyography (sEMG) for monitoring muscle activity and movement intention.

Purpose of the Study:

• To propose a novel combined scheme using Gray Relational Analysis (GRA) and an improved Cuckoo Search optimized Support Vector Regression (ICS-SVR) model to estimate knee joint extension force.
• To evaluate the accuracy and superiority of the proposed MMG-based method compared to existing regression models for muscle force estimation.

Main Methods:

• Extraction of comprehensive muscle activity features from MMG signals, including time-domain, frequency-domain, time-frequency-domain, and nonlinear dynamics features.
• Application of GRA to identify and select features highly correlated with knee joint extension force.
• Development and implementation of the ICS-SVR model, integrating selected features for accurate muscle force estimation.

Main Results:

• The combined GRA and ICS-SVR scheme demonstrated superior performance in estimating knee joint extension force compared to other regression models.
• The proposed method achieved higher estimation accuracy for muscle force, indicating its effectiveness.
• Experimental results validated the model's capability to meet the demands of muscle force estimation for advanced rehabilitation devices.

Conclusions:

• The proposed MMG-based GRA and ICS-SVR scheme provides a highly accurate and noninvasive method for estimating knee joint extension force.
• This approach holds significant potential for applications in rehabilitation devices, powered prostheses, and other assistive technologies requiring precise muscle force monitoring.