Adaptive neuro-fuzzy logic analysis based on myoelectric signals for multifunction prosthesis control

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Signal Processing

Background:

• Myoelectric signals reflect intended muscle contractions and movement.
• Amputees can generate repeatable myoelectric signals.
• Surface electromyography (sEMG) offers a non-invasive method for capturing these signals.

Purpose of the Study:

• To investigate the classification of five distinct arm movements using forearm sEMG signals.
• To evaluate the efficacy of an adaptive neuro-fuzzy inference system (ANFIS) for movement recognition.
• To determine the feasibility of using a minimal number of electrodes for prosthetic control.

Main Methods:

• Utilized forearm surface electromyography (sEMG) signals from three subjects.
• Employed three pairs of surface electrodes placed strategically on the forearm.
• Classified five distinct arm movements using an adaptive neuro-fuzzy inference system (ANFIS).

Main Results:

• The ANFIS achieved an average classification accuracy of 86-98% for the five motion classes.
• Demonstrated successful recognition of intended movements based on sEMG patterns.
• Validated the effectiveness of the chosen electrode placement and ANFIS model.

Conclusions:

• Forearm sEMG signals can be reliably used to classify multiple arm movements.
• The ANFIS model provides an effective approach for interpreting myoelectric control signals.
• This method shows promise for advanced prosthetic limb control in amputees.