Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back Dynamic gesture recognition system based on wearable device V.A. BEREZHNOY D.A. YUDIN Paper considers the task of gesture recognition which is important to build human-machine interfaces for hardware and software control of computers, robots, to improve the communication ability of people with hearing and speech defects. The authors have developed a system for capturing, evaluating, classifying dynamic human hand gestures. It is built on the basis of the wearable device. Gesture recognition is carried out by using etalon representation of gestures. Also it is based on C-means clustering algorithm, the usage of fuzzy finite state machines and grammars with the construction of gesture fuzzy membership functions. Article shows an approach to building models of dynamic gestures and an example of constructing a model of the etalon gesture graph. Wireless wearable device and software for it are developed for implementing gesture recognition. Article also shows the structure of a wearable device, its functionality and appearance of its hardware component based on Arduino Nano. System testing was made for nine different dynamic gestures. It showed acceptable accuracy of their recognition and the ability of the system working in real time. Article also represents the prospects for the further development and application of the developed system. Key words: dynamic gestures analysis, clustering, C-means method, clear finite state machines, fuzzy finite state machines, fuzzy membership function, gesture recognition system, wearable device. Contacts: E-mail: yuddim@yandex.ru Pp. 57-65.