Researchers achieve an 86% response accuracy rate for a hand orthosis for stroke rehabilitation, using surface electromyography signals

26 Mar 2026

Loss of control in gripping with the hand is a possible long-term effect of stroke. Recovery from this is possible through rehabilitation. Robotic assistive technology is being explored and used for such rehabilitation. Using surface electromyography (sEMG) signals from the arm, the recovering stroke patient can control the robotic assistive device for rehabilitation. This is the myoelectric hand orthosis.

Based on studies conducted in the early 2000s, patients who lose or experience impaired motor function in the upper limb are more likely to recover if the therapy includes active patient participation on a neurological level, rather than passively following a predetermined set of movements with an assistive device. Thus, robotic assistive technology needs to function on a stimulus or control initiated by the patient.

Surface electromyography is a non-invasive technique that detects muscle activity and translates it as an electrical signal that can be measured in the frequency spectrum, amplitude, or the action potential. Using sEMG has proven to be an accurate control signal in the context of robotic assistive technology in rehabilitation, by using an array of electrodes to cover the major muscle groups involved in the movement being allowed and aided by the assistive technology.

This work explored the potential of using the magnitude of the sEMG signal to control the force exerted by the hand orthosis and to make it proportional to the user’s intended force. Testing results showed a linear, directly proportional relationship between the user’s applied force and the magnitude of the electromyography signal, enabling this control signal to be used in force-control applications of the prototype. The prototype recorded a maximum force of 25.48 N, which it can supply to assist the user in gripping motions. It also achieved 86% response accuracy for hand grip and release motions.

Authors: Micah Angelo R. Bacani (Electrical and Electronics Engineering Institute, College of Engineering, University of the Philippines Diliman) and Manuel Ramos Jr. (Electrical and Electronics Engineering Institute, College of Engineering, University of the Philippines Diliman)

Read the full paper: https://dl.acm.org/doi/pdf/10.1145/3685073

Researchers achieve an 86% response accuracy rate for a hand orthosis for stroke rehabilitation, using surface electromyography signals

Loss of control in gripping with the hand is a possible long-term effect of stroke. Recovery from this is possible through rehabilitation. Robotic assistive technology is being explored and used for such rehabilitation. Using surface electromyography (sEMG) signals from the arm, the recovering stroke patient can control the robotic assistive device for rehabilitation. This is the myoelectric hand orthosis.

Based on studies conducted in the early 2000s, patients who lose or experience impaired motor function in the upper limb are more likely to recover if the therapy includes active patient participation on a neurological level, rather than passively following a predetermined set of movements with an assistive device. Thus, robotic assistive technology needs to function on a stimulus or control initiated by the patient.

Surface electromyography is a non-invasive technique that detects muscle activity and translates it as an electrical signal that can be measured in the frequency spectrum, amplitude, or the action potential. Using sEMG has proven to be an accurate control signal in the context of robotic assistive technology in rehabilitation, by using an array of electrodes to cover the major muscle groups involved in the movement being allowed and aided by the assistive technology.

This work explored the potential of using the magnitude of the sEMG signal to control the force exerted by the hand orthosis and to make it proportional to the user’s intended force. Testing results showed a linear, directly proportional relationship between the user’s applied force and the magnitude of the electromyography signal, enabling this control signal to be used in force-control applications of the prototype. The prototype recorded a maximum force of 25.48 N, which it can supply to assist the user in gripping motions. It also achieved 86% response accuracy for hand grip and release motions.

Authors: Micah Angelo R. Bacani (Electrical and Electronics Engineering Institute, College of Engineering, University of the Philippines Diliman) and Manuel Ramos Jr. (Electrical and Electronics Engineering Institute, College of Engineering, University of the Philippines Diliman)

Read the full paper: https://dl.acm.org/doi/pdf/10.1145/3685073