Elektromyografické ovládání robotické ruky
Loading...
Date
2021-10-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
V nedávné době se některé firmy v~oblasti protetiky začaly ubírat směrem zjednodušení a~zefektivnění procesu výroby protéz nebo jejich komponent. Jednou z hlavních hnacích sil, které umožnily vznik tohoto trendu, byl bezpochyby rapidní rozvoj technologií 3D~tisku, které například již umožňují tisknout díly se stejnými vlastnostmi, jako kdyby byly vyrobeny metodou vstřikování plastu do~formy, nebo třeba tisk kovových částí. V návaznosti na rozvoj těchto technologií se pak vyskytly i projekty, kdy uživatel nemusí kvůli výrobě lůžka pro protézu dojíždět za odborníkem, nýbrž mu je~zaslán 3D~skener, kterým si~v~pohodlí domova naskenuje pahýl a~následně zpět zašle přístroj s~daty, na~základě kterých mu je~vyrobeno lůžko na~míru. Tato práce si klade ze cíl dále sledovat tento směr vývoje, jež je~zaměřen na ulehčení koncovému uživateli. Za tímto účelem byl navržen systém vyhodnocující provedená gesta na základě snímaných SEMG signálů, jejichž podoba hodnoty se mění v průběhu dne, který může být kdykoli během dne rychle přetrénován, a~to~i~když uživatel zrovna nesedí u počítače. Tento způsob návrhu umožňuje snadné obnovení vysoké klasifikační přesnosti systému.
Recently, some companies in the field of prosthetics have shifted towards making the process of manufacturing the prosthetic devices and their components more effective and streamlined. Among the driving forces that had made this shift possible, one of the more cruicial ones was without a doubt the advent of 3D printing technologies, which has for instance allowed for printing components whose mechanical properties are on par with their injection moulded counterparts, or printing parts with metal. Coinciding with these developments, approaches which aim to make the process of fitting the prosthetic device more user friendy, such as sending a 3D scanning device to a patient's house, where the patient scans their stump and returns the device along with the scanned data to the sender in order for them to manufacture a custom socket, eliminating the need for the patient to visit a professional, have started to~emerge. The aim of this Master's thesis is to expand upon forementioned trend. The thesis presents a system that utilizes SEMG data, values of which fluctuate throughout the course of a day for respective gestures, that can be efficiently retrained anywhere, eliminating the need for the user to be by a computer at~the time of~training. This approach grants the user the ability to~achive higher classification procision throughout the day, wherever they may be.
Recently, some companies in the field of prosthetics have shifted towards making the process of manufacturing the prosthetic devices and their components more effective and streamlined. Among the driving forces that had made this shift possible, one of the more cruicial ones was without a doubt the advent of 3D printing technologies, which has for instance allowed for printing components whose mechanical properties are on par with their injection moulded counterparts, or printing parts with metal. Coinciding with these developments, approaches which aim to make the process of fitting the prosthetic device more user friendy, such as sending a 3D scanning device to a patient's house, where the patient scans their stump and returns the device along with the scanned data to the sender in order for them to manufacture a custom socket, eliminating the need for the patient to visit a professional, have started to~emerge. The aim of this Master's thesis is to expand upon forementioned trend. The thesis presents a system that utilizes SEMG data, values of which fluctuate throughout the course of a day for respective gestures, that can be efficiently retrained anywhere, eliminating the need for the user to be by a computer at~the time of~training. This approach grants the user the ability to~achive higher classification procision throughout the day, wherever they may be.
Description
Subject(s)
SEMG, ML, SVM, ESP32, Myo~Armband, SEMG, ML, SVM, ESP32, Myo~Armband