Multi-channel Control System Using Force Myography

The creation of a bionic system capable of almost completely restoring the functionality of a lost lamb is the main task of prosthetics. This article discusses the need to develop a control system capable of classifying a large number of gestures with high accuracy and acceptable delay. This article discusses the use of power myography to build a control system for bionic devices. The object of the study was the upper third of the forearm with the most commonly used movements in everyday life: flexion, extension, retraction, reduction, capture, pronation, supination. The registration of power myography signals was carried out using four resistive pressure sensors, which were attached using a special bracelet. Ten healthy volunteers participated in the study. Based on the data obtained, it was possible to classify actions using machine learning methods with an average accuracy of $81.4 \%$. The signals recorded using this system can be used in the development of prosthetic control systems, in the development of virtual and additional reality interfaces, in sports science.