Abstract
Sonomyography (SMG) is a method of controlling upper-limb prostheses through an innovative human-machine interface by monitoring forearm muscle activity through ultrasonic imaging. Over the past two decades, SMG has shown promise, achieving over 90% accuracy in classifying hand gestures when combined with artificial intelligence, making it a viable alternative to electromyography (EMG). However, up to now, there are few reports of a system integrating SMG together with a prosthesis for testing on amputee subjects to demonstrate its capability in relation to daily activities. In this study, we developed a highly efficient human-machine interface algorithm for controlling a prosthetic hand with 6-DOF using a wireless and wearable ultrasound imaging probe. We first evaluated the accuracy of our model in classifying nine different hand gestures to determine its reliability and precision. The results from the offline study, which included ten healthy participants, indicated that nine different hand gestures could be classified with a success rate of 100%. Additionally, the developed controlling system was tested in real-time experiments on two amputees, using a variety of hand function test kits. The results from the hand function tests confirmed that the prosthesis, controlled by the SMG system, could assist amputees in performing a variety of hand movements needed in daily activities.