PPGINF PÓS-GRADUAÇÃO EM ENGENHARIA DA INFORMAÇÃO FUNDAÇÃO UNIVERSIDADE FEDERAL DO ABC Teléfono/Ramal: No informado http://propg.ufabc.edu.br/ppginfo

Banca de QUALIFICAÇÃO: FELIPE SADAMI OIWA DA COSTA

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE : FELIPE SADAMI OIWA DA COSTA
DATA : 15/12/2021
HORA: 14:00
LOCAL: Link remoto
TÍTULO:

sEMG Signal Acquisition and Pattern Extraction via Automatic Synthesis of Dynamical Systems for Bionic Hands Prosthesis


PÁGINAS: 70
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Elétrica
SUBÁREA: Medidas Elétricas, Magnéticas e Eletrônicas; Instrumentação
ESPECIALIDADE: Instrumentação Eletrônica
RESUMO:

The lack or loss of a limb by an Individual can enormously impact their daily life’s mobility. Millions of people suffer from amputations, congenital diseases, and other physical disabilities around the world. The development of prostheses came as a solution for replacing individuals’ absent limb’s function and aesthetics using artificial components. Initially, prostheses only had an aesthetic role. However, the technological evolution allowed the development of new prostheses with advanced mechatronics and intelligent control, which led to a new field of study known as Robotic Prosthetics. The idea of developing an artificial robotic limb that could fully replace the human one in terms of functionality and aesthetics has shown to be a complex task that raised many others branches in the study of Robotic Prosthetics. There are two main groups of studies in Robotic Prosthetics, those related to the Architecture of the prosthesis and those focused on Biosignal Acquisition, Processing, and Control. The architecture of a human arm or hand is unique, and trying to mimic or reproduce it, even using the most advanced components, is still a challenging job. Studies of the Architecture of Prosthesis involve better forms of actuation, new materials for bones and skins aesthetics, sensors, concepts for improving comfort and performance, and low-cost prostheses. On the other side, human biosignal studies try to understand and translate it in control command for Robotic Prosthesis. Studies involving biosignals, such as EEG and EMG, are complex, require biological comprehension, and demand deep knowledge of Signal Processing, Hardware, and Software design. The main topics studying biosignals involve Signal Acquisition, Dimensionality Reduction, Feature Extraction, and Decoding/Control. In Signal Processing, the decoding/control of a signal can be divided into two approaches: Pattern and Non-Pattern Recognition. Pattern Recognition uses classification-based techniques for decoding signals to control commands. Basically, in the Pattern Recognition approach, the information inside a signal is extracted in features (Feature Extraction) that the classifier uses to identify a respective class. A class corresponds to a discrete predetermined movement. Literature shows that Patter Recognition-based studies reached classification accuracies of up to 99%. However, the limitations in the number of classes, the lack of continuous, independent, and proportional control of the arm, hand, and fingers have not yet allowed the development of a fully functional Robotic Prosthesis. On the other hand, many studies have been developed using other than classification-based approaches, also called Non-Pattern Recognition, to overcome the aforementioned problems. In this direction, this work proposes using Automatic Synthesis of Non-Linear Dynamical Systems to find a suitable model that extracts and translates the information inside sEMG signals to fingers trajectory in 3D space. It is built a self-constructive model using NARX architecture in conjunction with Metaheuristics or Bioinspired Optimization Algorithms such as Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). A custom sEMG Signal Acquisition System was constructed due to the specific requirements of the proposal.


MEMBROS DA BANCA:
Presidente - Interno ao Programa - 1205456 - LUIZ ALBERTO LUZ DE ALMEIDA
Membro Titular - Examinador(a) Interno ao Programa - 1761107 - RICARDO SUYAMA
Membro Titular - Examinador(a) Externo ao Programa - 1672981 - FRANCISCO JAVIER ROPERO PELAEZ
Membro Suplente - Examinador(a) Externo ao Programa - 2078059 - LUIZ ANTONIO CELIBERTO JUNIOR
Notícia cadastrada em: 29/11/2021 13:13
SIGAA | UFABC - Núcleo de Tecnologia da Informação - ||||| | Copyright © 2006-2024 - UFRN - sigaa-2.ufabc.int.br.sigaa-2-prod