Home / Current Issue / Paper 1712170
AI-Enabled Wearable System for Real - Time Prediction of Asthma Exacerbation
Subject area: Science,Engineering and Technology · Area of research: Wearable AI for Early Asthma Attack Prediction
DOI: https://doi.org/10.64388/IREV9I5-1712170
Abstract
This research focuses on developing a wearable AI-based system for early detection of asthma attacks through real-time physiological monitoring. The device uses a single photoplethysmography (PPG) sensor with an embedded machine learning model to enable on-device prediction without relying on cloud servers. Two systems were compared: a traditional cloud-based multi-sensor IoT model and the proposed single-sensor AI device, using 3000 samples of heart rate, oxygen saturation, and respiratory rate. The Random Forest algorithm was optimized for microcontroller use, and performance metrics were analyzed using SPSS tools. Results showed that the proposed system achieved 95.2% accuracy, 93.8% sensitivity, and 96.4% specificity, along with reduced latency and power consumption. The findings demonstrate that on-device AI enhances efficiency, reliability, and speed, making the system suitable for continuous respiratory health monitoring and early asthma prevention.
How to cite this paper
@article{1712170,
author = {Dr. V. Priya, M. Aruna, S. Anitha, S. Jenith Anisha},
title = {AI-Enabled Wearable System for Real - Time Prediction of Asthma Exacerbation},
journal = {Iconic Research And Engineering Journals},
year = {2025},
volume = {9},
number = {5},
pages = {1368-1374},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1712170.pdf},
abstract = {This research focuses on developing a wearable AI-based system for early detection of asthma attacks through real-time physiological monitoring. The device uses a single photoplethysmography (PPG) sensor with an embedded machine learning model to enable on-device prediction without relying on cloud servers. Two systems were compared: a traditional cloud-based multi-sensor IoT model and the proposed single-sensor AI device, using 3000 samples of heart rate, oxygen saturation, and respiratory rate. The Random Forest algorithm was optimized for microcontroller use, and performance metrics were analyzed using SPSS tools. Results showed that the proposed system achieved 95.2% accuracy, 93.8% sensitivity, and 96.4% specificity, along with reduced latency and power consumption. The findings demonstrate that on-device AI enhances efficiency, reliability, and speed, making the system suitable for continuous respiratory health monitoring and early asthma prevention.},
month = {November},
doi = {https://doi.org/10.64388/IREV9I5-1712170}
}