Key points are not available for this paper at this time.
Abstract This review examines 49 articles published recently that include artificial intelligence (AI) as a method of incorporating into the design of prosthetic hands, primarily using electromyography (EMG) signals for control. The review identified that the use of deep learning, neural network techniques, and EMG significantly improved grasp accuracy, response time, and functional performance, bringing prosthetic hands closer to approximating naturalistic user interaction. Nevertheless, some significant challenges still exist, including high power consumption, inaccurate signal interpretation, processing delays, high cost, and limited reliability in real-world contexts. Therefore, future studies should include energy-efficient adaptive algorithms and multiple sources of information (e.g., tactile and optical/visual sensors) to create more intelligent, adaptable, and user-friendly robotic prosthetic limbs. The ultimate goal is to enhance accuracy, independence, and quality of life.
Amjed Abdulameer Al-Musawi (Sat,) studied this question.