Proportional and Simultaneous Real-Time Control of the Full Human Hand From High-Density Electromyography

• DOI: 10.36227/techrxiv.21904335
• Publication Date: 2023-01-24T16:26:35Z
• Authors: Raul Sîmpetru; Michael März; Alessandro Del Ve...
• Keywords: Motor Control; Smoothing

<p>Surface electromyography (sEMG) is a non-invasive technique that measures the electrical activity generated by muscles using sensors placed on skin. It has been widely used in field of prosthetics and other assistive systems because physiological connection between muscle movement dynamics. However, most existing sEMG-based decoding algorithms show limited number detectable degrees freedom can be proportionally simultaneously controlled real-time, which limits use EMG wide range applications, including consumer-level applications (e.g., human/machine interfacing). In this work, we surpass current state art developing new deep learning method decode map electrophysiological forearm into proportional simultaneous control > 20 human hand with real-time resolution latency within neuromuscular delays (< 50 ms). We recorded kinematics during grasping, pinching, individual digit movements three gestures at slow (0.5 Hz) fast (0.75 speeds healthy participants.</p> <p>We demonstrate our neural network predict constant 32 predictions per second. To achieve this, employed transfer created prediction smoothing algorithm for output reconstructed full geometry three-dimensional Cartesian space real-time. Our results high-density signals from contain almost all information needed to hand. The proposed capability predicting an unprecedented way immediate translational impact subjects motor impairments. </p>