Smart Health App : monitorización de movimientos para pacientes de ictus utilizando Java

Roy Cubillo, Adrian (2018). Smart Health App : monitorización de movimientos para pacientes de ictus utilizando Java. Thesis (Master thesis), E.T.S.I. Telecomunicación (UPM).

Description

Title: Smart Health App : monitorización de movimientos para pacientes de ictus utilizando Java
Author/s:
  • Roy Cubillo, Adrian
Contributor/s:
  • Jingyue, Li
Item Type: Thesis (Master thesis)
Masters title: Ingeniería de Telecomunicación
Date: 2018
Subjects:
Freetext Keywords: Stroke patients; Android application; Human activity recognition took; Machine learning algorithms
Faculty: E.T.S.I. Telecomunicación (UPM)
Department: Otro
Creative Commons Licenses: Recognition - No derivative works - Non commercial

Full text

[img]
Preview
PDF - Requires a PDF viewer, such as GSview, Xpdf or Adobe Acrobat Reader
Download (1MB) | Preview

Abstract

Stroke patients suffer from different temporal and/or permanent consequences that include movement disabilities. They are usually treated with physical rehabilitation which can compound different exercises, some of them performed by specialists or physical therapists and others involving exercises performed by the patient on his/her own. Therefore, it is important for the doctor (and also for the patient) tracking the performance of the exercises done by the patient on his/her daily life. This information can help doctors and physiotherapists catch information about progress and movements that sometimes are not possible to get by just talking with the patient. The information may include the activities the patient performs such as running, walking, standing, etc... and the time the patient spends on them. Doctors can analyze the activities and times and extract results which can help them on seeing what needs to be improved by the patient, or what the patient lacks of, making afterwards a more specialized treatment adapting it to the specific patient thanks to the collected data. Smart Health App project aims to create an android application which is able to classify the activities performed by a patient, analyzing the accelerometer data given by the mobile phone sensors. The application is inspired on previously made Human Activity Recognition (HAR) systems such as Activity Recognition for Stroke Patients[16], trying to improve the different lacks of them. The application is able to record users’ movements save them and label them on the training process. Afterwards, it is able to classify the movements data (accelerometer information) into specific activities relying on random forest machine learning algorithm for the human activity recognition process. The main improvement is making all the process straightforward and on the same application so that users and doctor can in a relatively few time obtain the output classification results. This project proved that making an android based Human Activity Recognition tool is possible classifying movements into activities with a good accuracy, which is over 90% of correctly classified instances. That results were obtained with themobile phone’s only accelerometer sensor, proving then that it is possible to use that only one accelerometer sensor to correctly classify movements of stroke patients. With the mobile phone as only sensor device, it was important testing what are the best placements for the mobile phone while the user performs the activity. Two different placements were compared, hand and arm as they both were practical places to hold a mobile phone on. Arm placement provided better results. And last, some other machine learning algorithms were tested to see whether random forest is the best algorithm on stroke HAR.

More information

Item ID: 52872
DC Identifier: http://oa.upm.es/52872/
OAI Identifier: oai:oa.upm.es:52872
Deposited by: Biblioteca ETSI Telecomunicación
Deposited on: 30 Oct 2018 11:09
Last Modified: 31 Oct 2018 09:07
  • Logo InvestigaM (UPM)
  • Logo GEOUP4
  • Logo Open Access
  • Open Access
  • Logo Sherpa/Romeo
    Check whether the anglo-saxon journal in which you have published an article allows you to also publish it under open access.
  • Logo Dulcinea
    Check whether the spanish journal in which you have published an article allows you to also publish it under open access.
  • Logo de Recolecta
  • Logo del Observatorio I+D+i UPM
  • Logo de OpenCourseWare UPM