| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 8.34 MB | Adobe PDF |
Authors
Abstract(s)
O envelhecimento da população é uma tendência confirmada pelos últimos estudos realizados à população, tanto a nível nacional como internacional. Assim,
garantir qualidade de vida e uma adequada monitorização, nas mais diversas dimensões dos idosos nas suas habitações, revela-se uma prioridade e, naturalmente, uma
área de estudo relevante. A identificação de padrões de movimento e o treino de um
algoritmo de machine learning para a deteção de outliers, representam os objetivos
principais do trabalho.
A monitorização de pessoas está frequentemente associada à perda de privacidade e autonomia. De forma a evitar estas situações, propusemo-nos explorar
a capacidade de reutilizar soluções de IoT e sensorização existentes na habitação.
Pretende-se extrair das soluções de smart homes, o máximo de informação possível
para, numa primeira fase, estudar os padrões de movimento dos habitantes e, dessa
forma, alavancar um conjunto de ações que potenciem o seu bem estar e monitorização. Para atingir os objetivos propostos no projeto, foram implementadas várias
abordagens para o estudo de padrões, a análise da distribuição dos registos, o cálculo dos percentis e o cálculo da probabilidade da ocupação. Implementou-se um
algoritmo de machine learning para proceder à identificação de outliers nos dados
extraídos dos sensores da habitação.
ABSTRACT: The aging of the population is a trend that has been confirmed by national and international studies. Identifying and studying the movement patterns aswell as the training of a machine learning model to identify outliers, were the main goals of this project. Ensuring quality of life and a adequated monitorization of the elderly population in their homes has proven to be a priority and a relevant area of research. The monitorization of people is frequently associated with privacy and autonomy issues. In order to be able to avoid this, existing IoT and Smart Home Solutions was explored. The aim was to extract from Smart Home solutions as much information as possible so that, movement patterns could be identified and studied and thereby leverage a set of improvements of the monitoring process. To achieve these improvements a machine learning algorithm was implemented, so that, outliers could be identified in the data that was extracted from the sensores.
ABSTRACT: The aging of the population is a trend that has been confirmed by national and international studies. Identifying and studying the movement patterns aswell as the training of a machine learning model to identify outliers, were the main goals of this project. Ensuring quality of life and a adequated monitorization of the elderly population in their homes has proven to be a priority and a relevant area of research. The monitorization of people is frequently associated with privacy and autonomy issues. In order to be able to avoid this, existing IoT and Smart Home Solutions was explored. The aim was to extract from Smart Home solutions as much information as possible so that, movement patterns could be identified and studied and thereby leverage a set of improvements of the monitoring process. To achieve these improvements a machine learning algorithm was implemented, so that, outliers could be identified in the data that was extracted from the sensores.
Description
Keywords
Geriatria Internet of Things (IoT) Envelhecimento Sensores Smart Home Monitorização Outliers