Grip Force and 3D Push-Pull Force Estimation Based on sEMG and GRNN.

Area of Science:

• Biomedical Engineering
• Robotics
• Neuroscience

Background:

• Dexterous control of prosthetic hands relies on accurate estimation of grip and 3D forces.
• Electromyogram (EMG) signals offer a promising avenue for real-time force estimation.
• Existing methods may lack the precision required for advanced prosthetic functionality.

Purpose of the Study:

• To propose and validate a novel method for estimating grip and 3D push-pull forces using surface EMG (sEMG) signals.
• To develop an intelligent EMG prosthetic hand system capable of precise force control.
• To enhance the functionality and user experience of EMG-controlled prosthetic limbs.

Main Methods:

• Utilized eight-channel surface EMG (sEMG) data acquired from arm skin.
• Employed a Generalized Regression Neural Network (GRNN) for force estimation.
• Measured actual hand forces using grip and 3D force sensors concurrently with sEMG acquisition.

Main Results:

• The GRNN model effectively estimated human hand grip and 3D push-pull forces from sEMG signals.
• High correlation coefficients were observed between actual and estimated forces.
• Analysis of variance (ANOVA) confirmed the statistical significance of the estimation accuracy.

Conclusions:

• The proposed sEMG and GRNN-based method accurately estimates action forces, meeting the requirements for intelligent prosthetic hand control.
• This approach holds significant potential for improving the dexterity and responsiveness of EMG prosthetic hands.
• The findings pave the way for more intuitive and effective human-machine interfaces in prosthetics.