Myoelectric control algorithm for robot-assisted therapy: a hardware-in-the-loop simulation study.

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Kinesiology

Background:

• Anterior Cruciate Ligament (ACL) injuries often result from direct knee trauma.
• Robot-assisted therapy (RAT) enhances recovery post-ACL reconstruction, but its application is limited to later rehabilitation phases.
• There is a need for assistive technologies in earlier ACL rehabilitation stages, prior to advanced exercises like walking.

Purpose of the Study:

• To propose and evaluate a myoelectric control (MEC) algorithm for the "Nukawa" robot-assisted rehabilitation system.
• To enable knee movement assistance during early-phase ACL rehabilitation exercises, such as active-assisted extension.
• To integrate surface electromyography (sEMG) for detecting knee joint motion intention.

Main Methods:

• Developed a surface electromyography (sEMG) signal processing algorithm to detect knee joint motion intention.
• Utilized a hardware-in-the-loop simulation, combining sEMG processing and movement control algorithms, to test the MEC algorithm.
• Conducted experiments with 17 healthy subjects performing 12 ACL rehabilitation exercises.

Main Results:

• The MEC algorithm achieved 100% accuracy in detecting the orientation of knee extension and flexion intention.
• Movement intention intensity was detected comparably to maximum voluntary contraction (MVC) in 94% of extension and 59% of flexion exercises.
• Maximum position mean absolute errors for hip, knee, and ankle joints were reported, indicating system precision.

Conclusions:

• The MEC algorithm effectively detects knee movement intention intensity and orientation, comparable to MVC.
• The system requires no prior subject training or additional torque sensors.
• The algorithm controls the "Nukawa" robot's knee joint speed to facilitate active-assisted exercises in ACL rehabilitation.