Low Complex CORDIC-based Hand Movement Recognition Design Methodology for Rehabilitation and Prosthetic Applications.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Rehabilitation Technology

Background:

• Electromyography (EMG) signal analysis is vital for rehabilitation and prosthetic applications, particularly for neuromuscular disorders.
• Existing EMG-based hand movement recognition systems face limitations due to high computational complexity and power consumption, hindering their use in resource-constrained environments.
• Current methods often involve computationally intensive algorithms like Ensemble Empirical Mode Decomposition (EEMD), Fast Independent Component Analysis (FastICA), and Linear Discriminant Analysis (LDA), which are difficult to implement on low-complexity hardware.

Purpose of the Study:

• To develop a low-complexity hand movement recognition methodology for EMG signals suitable for resource-constrained rehabilitation applications.
• To replace the computationally intensive Linear Discriminant Analysis (LDA) classifier with a more efficient K-Means clustering algorithm.
• To further reduce computational complexity by employing a CORDIC-based K-Means clustering approach.

Main Methods:

• A novel design methodology for hand movement recognition using single-channel EMG data was proposed.
• Linear Discriminant Analysis (LDA) classification was replaced with K-Means clustering to reduce computational load.
• A CORDIC-based K-Means clustering algorithm was implemented to optimize system complexity.

Main Results:

• The proposed CORDIC-based K-Means clustering method achieved a 99.77% reduction in complexity compared to conventional LDA-based classification.
• The new methodology demonstrated a 1.28% increase in accuracy for classifying seven hand movements using single-channel EMG data.
• The system is designed for low-complexity and efficient operation, addressing power consumption and on-chip area constraints.

Conclusions:

• The developed CORDIC-based K-Means clustering approach offers a significantly more efficient and accurate solution for EMG-based hand movement recognition.
• This methodology is well-suited for resource-constrained rehabilitation applications, enabling the development of more accessible and effective assistive technologies.
• The findings highlight the potential of low-complexity algorithms in advancing wearable and implantable biomedical devices.