Robótica de assistência a idosos: apoio à toma de medicamentos
Data
2018-11-28
Título da revista
ISSN da revista
Título do Volume
Editora
Resumo
O envelhecimento da população mundial e a correspondente diminuição de população
jovem ativa nos países desenvolvidos está a provocar uma grande pressão sobre os recursos
humanos especializados na prestação de cuidados necessários às pessoas idosas.
Frequentemente, os idosos possuem um regime de medicação vasto e complexo. Aliado a
isto, os constantes avanços na tecnologia e neste caso específico, na robótica, fazem com que
os robôs surjam como uma alternativa para preencher a lacuna existente entre a necessidade e
o fornecimento de cuidados de saúde a idosos.
Embora existam diversas soluções tecnológicas que visam o auxílio à toma de medicação,
todas elas são limitadas na interação com o utilizador, nomeadamente incapazes de irem ao
encontro do idoso e de lhe entregar o medicamento e a água no horário devido.
Nesta dissertação é proposto um sistema robótico que visa o auxílio a idosos, ajudando-os
no cumprimento do regime de medicação a que estejam sujeitos. O robô usa o seu sistema de
locomoção para se movimentar até ao idoso e através de visão por computador deteta a
embalagem do medicamento e identifica a pessoa que o deve tomar no horário correto. Para a
realização da tarefa foi desenvolvida uma aplicação suportada por uma base de dados com
informações sobre os idosos da habitação, os medicamentos que têm prescritos e os respetivos
horários de toma.
O trabalho experimental foi efetuado com o robô NAO, da empresa SoftBank Robotics,
recorrendo a ferramentas de desenvolvimento como o MySQL, Python e OpenCV. A
identificação facial dos idosos e o reconhecimento dos medicamentos é efetuado através
algoritmos de visão por computador que processam as imagens adquiridas pela câmara do robô.
Com o objetivo de avaliar o desempenho dos algoritmos de visão por computador, foram
efetuados testes de reconhecimento de objetos, deteção facial e reconhecimento facial
utilizando bases de dados públicas. Os testes possibilitaram obter métricas qualitativas do
desempenho dos algoritmos utilizados.
Para demonstração do conceito foi feito um teste num cenário simples que recria o
ambiente de uma habitação com idosos que são assistidos pelo robô na toma de medicamentos
The ageing of the world's population and the corresponding decrease in the active young population in developed countries, are putting a great strain on human resources specialized in the care of the elderly population. Often, the elderly have a vast and complex medication regimen. Alongside this, the constant advances in technology and, in this specific case, in robotics, make the robots appear as an alternative to fill the gap between the need and supply of health care for the elderly population. Although there are several technological solutions that provide help in taking their medications, they are all limited in the interaction with the user. In this case, they are unable to walk to the elderly and deliver the medication and water in time. In this dissertation it is proposed a robotic system that provides assistance to the elderly, helping them to accomplish the medication regimen to which they are subject. The robot uses its locomotion system to move to the elderly and through computer vision, detects the packaging of the medicine and identifies the person who should take it at the correct time. For the accomplishment of the task an application was developed supported by a database with information about the elderly of the dwelling, the medicines that they have been prescribed and the respective time of intake. The experimental work was done with the robot NAO, of the company SoftBank Robotics, using development tools like MySQL, Python and OpenCV. Facial recognition of the elderly and the identification of the right medications are performed through computer vision algorithms that process the images acquired by the robot's camera. In order to evaluate the performance of the computer vision algorithms, object recognition, facial detection and facial recognition, tests were performed using public databases. The tests enable the acquisition of qualitative metrics of the performance accomplished by the algorithms used. For demonstration of the concept a test was done in a simple scenario that recreates the environment of a dwelling with elderly people who are assisted by the robot in the intake of the medication.
The ageing of the world's population and the corresponding decrease in the active young population in developed countries, are putting a great strain on human resources specialized in the care of the elderly population. Often, the elderly have a vast and complex medication regimen. Alongside this, the constant advances in technology and, in this specific case, in robotics, make the robots appear as an alternative to fill the gap between the need and supply of health care for the elderly population. Although there are several technological solutions that provide help in taking their medications, they are all limited in the interaction with the user. In this case, they are unable to walk to the elderly and deliver the medication and water in time. In this dissertation it is proposed a robotic system that provides assistance to the elderly, helping them to accomplish the medication regimen to which they are subject. The robot uses its locomotion system to move to the elderly and through computer vision, detects the packaging of the medicine and identifies the person who should take it at the correct time. For the accomplishment of the task an application was developed supported by a database with information about the elderly of the dwelling, the medicines that they have been prescribed and the respective time of intake. The experimental work was done with the robot NAO, of the company SoftBank Robotics, using development tools like MySQL, Python and OpenCV. Facial recognition of the elderly and the identification of the right medications are performed through computer vision algorithms that process the images acquired by the robot's camera. In order to evaluate the performance of the computer vision algorithms, object recognition, facial detection and facial recognition, tests were performed using public databases. The tests enable the acquisition of qualitative metrics of the performance accomplished by the algorithms used. For demonstration of the concept a test was done in a simple scenario that recreates the environment of a dwelling with elderly people who are assisted by the robot in the intake of the medication.
Descrição
Dissertação de Mestrado em Engenharia Informática apresentada à Universidade de Trás-os-Montes e Alto Douro
Palavras-chave
Robótica de assistência a idosos , Visão por computador , Reconhecimento de objetos , Reconhecimento facial