Wearable Ultrasound-based Decoding of Simultaneous Wrist/Hand Kinematics

The advancement in artificial hands challenges the development of human-machine interface. In this study, we present a wearable ultrasound based interface for decoding simultaneous and proportional (SP) wrist/hand kinematics. Different from traditional velocity-based SP control, this study focuses on position-based SP control that aligns with human proprioception. Results show that simultaneous wrist rotation and hand closing can be accurately decoded from ultrasound, albeit using a contralateral training scheme. Comparative analysis shows that the performance of ultrasound is significantly better than surface electromyogram. Moreover, it has potential to predict simultaneous multi-DoF wrist/hand kinematics using single-DoF training by a linearly enhanced training strategy. This is the first study to achieve simultaneous and proportional position control via ultrasound, paving the way for the intuitive interaction with future prosthetic devices.