A Method of Recognizing Finger Motion Using Wavelet Transform of Surface EMG Signal

Area of Science:

• Biomedical Engineering
• Signal Processing
• Neuroscience

Background:

• Surface electromyography (EMG) signals offer a promising, non-invasive method for detecting human movement intentions.
• Accurate interpretation of multi-channel EMG data is crucial for developing intuitive prosthetic limb control systems.

Purpose of the Study:

• To develop and validate a novel method for identifying finger motions using wavelet transform and Artificial Neural Networks (ANNs).
• To establish an alternative approach for utilizing surface EMG signals in controlling multi-fingered prosthetic hands.

Main Methods:

• Surface EMG signals from the upper arm were analyzed using multi-resolution wavelet transform.
• Key features, including variance, maximum, and mean absolute value of wavelet coefficients, were extracted to create a new feature space.
• The extracted feature values were input into an Artificial Neural Network (ANN) for finger motion identification.

Main Results:

• The wavelet transform effectively decomposed and analyzed the multi-channel EMG signals.
• A distinct feature space was successfully established using wavelet coefficients.
• Experimental results demonstrated the effectiveness of the proposed method in accurately identifying finger motions.

Conclusions:

• The presented wavelet transform-based EMG analysis method is effective for finger motion identification.
• This technique offers a viable alternative for EMG-based control of advanced prosthetic hands, enhancing user interaction and functionality.