Enhancing Water Safety in the Philippines: Early Detection of Drowning Patterns through Wearable Devices and Machine Learning Algorithms

Closed

Gadiel Nathan C. Arriesgado, Godwin S. Monserate, Bianca Isabel Y. Uytengsu

2025 Proceedings of the 22nd IEEE International Conference on Computer Applications, ICCA 2025 Conference paper Cited by 0 Quartile

Abstract

Drowning, described as the process of respiratory impairment/failure or suffocation due to submersion in water, remains one of the leading causes of preventable injury worldwide. Early intervention is crucial in reducing morbidity and mortality, as quick alerts and rapid response times are able to significantly lower the risk of permanent injury or death. The implementation of an automated system for detection and alerting of drowning patterns would then prove to have significant benefits to public health and safety. The researchers aim to create a system which can perform drowning detection through the use of technology that has become more ubiquitous in recent years, smartwatches. Smartwatches are compact mobile devices which are worn on the wrist that also may come with a variety of features and sensor technology. This study aims to both discuss and test the feasibility of using programmable smartwatches as a medium for reliable, affordable, and accurate detection of drowning patterns in real world scenarios. Collection of biometric data in real time will be used in tandem with machine learning algorithms such as the Local Outlier Factor (LOF) in order to detect anomalies in biometric data that may signal that a user is exhibiting symptoms of drowning. Furthermore, the study will also explore the challenges, limitations, and feasibility associated with the creation and implementation of the automated drowning detection system, including the need for accurate data collection, machine learning algorithm optimizations, and system validation and testing in real world scenarios. The potential applications of the proposed system may extend beyond individual user monitoring and may also have implications for improving current water safety measures among vulnerable groups, which could ultimately reduce the burden placed on water safety groups by drowning incidents. © 2025 IEEE.

Affiliations

University of San Carlos, Department of Computer Information, Science, and Mathematics, Cebu City, Philippines