Development of a Machine Learning Based in-Home Physical Activity Monitoring System Using Wrist Actigraphy and Real-Time Location System

Seyyed Mahdi Torabi, Mohammad Narimani, Edward Park

2024

Abstract

In this study, a multi-modal sensing approach was employed to enhance human activity recognition (HAR). The approach integrated data from a wearable wristband and a Real-Time Location System (RTLS) to perform physical posture classification (PPC) and indoor localization (IL). The performance of conventional machine learning techniques such as Logistic Regression (LR) and Long Short-Term Memory (LSTM) models were compared. The results demonstrated that LSTM models superior performance in terms of accuracy and robustness. The LSTM’s efficacy stems from its ability to capture temporal dependencies inherent in human activity data, making it suited for HAR tasks. Our findings underscored the benefits of employing a multi-modal, LSTM-based approach for enhancing HAR. The proposed approach increased the comprehensiveness of the HAR system. The proposed system holds potential for various in-home activity monitoring scenarios, suggesting promising implications for improving the quality of remote patient monitoring.

Paper Citation

in Harvard Style

Torabi S., Narimani M. and Park E. (2024). Development of a Machine Learning Based in-Home Physical Activity Monitoring System Using Wrist Actigraphy and Real-Time Location System. In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 1: BIODEVICES; ISBN 978-989-758-688-0, SciTePress, pages 135-141. DOI: 10.5220/0012368000003657

in Bibtex Style

@conference{biodevices24,
author={Seyyed Mahdi Torabi and Mohammad Narimani and Edward Park},
title={Development of a Machine Learning Based in-Home Physical Activity Monitoring System Using Wrist Actigraphy and Real-Time Location System},
booktitle={Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 1: BIODEVICES},
year={2024},
pages={135-141},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012368000003657},
isbn={978-989-758-688-0},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 1: BIODEVICES
TI - Development of a Machine Learning Based in-Home Physical Activity Monitoring System Using Wrist Actigraphy and Real-Time Location System
SN - 978-989-758-688-0
AU - Torabi S.
AU - Narimani M.
AU - Park E.
PY - 2024
SP - 135
EP - 141
DO - 10.5220/0012368000003657
PB - SciTePress