Towards On-Device Dehydration Monitoring Using Machine Learning from Wearable Device's Data.

Farida Sabry,Tamer Eltaras,Wadha Labda,Qutaibah Malluhi,Khawla Alzoubi,Fatima Hamza

doi:10.3390/s22051887

Farida Sabry, Tamer Eltaras + Show 4 more

Open Access

https://doi.org/10.3390/s22051887

Copy DOI

Abstract

With the ongoing advances in sensor technology and miniaturization of electronic chips, more applications are researched and developed for wearable devices. Hydration monitoring is among the problems that have been recently researched. Athletes, battlefield soldiers, workers in extreme weather conditions, people with adipsia who have no sensation of thirst, and elderly people who lost their ability to talk are among the main target users for this application. In this paper, we address the use of machine learning for hydration monitoring using data from wearable sensors: accelerometer, magnetometer, gyroscope, galvanic skin response sensor, photoplethysmography sensor, temperature, and barometric pressure sensor. These data, together with new features constructed to reflect the activity level, were integrated with personal features to predict the last drinking time of a person and alert the user when it exceeds a certain threshold. The results of applying different models are compared for model selection for on-device deployment optimization. The extra trees model achieved the least error for predicting unseen data; random forest came next with less training time, then the deep neural network with a small model size, which is preferred for wearable devices with limited memory. Embedded on-device testing is still needed to emphasize the results and test for power consumption.

Highlights

Following the tremendous technological advances in the design of systems on a chip (SoCs), the development and use of wearable devices have remarkably achieved high growth rates in the last few years
While persons normally feel thirst when they are going into dehydration to alert them to drink enough water to feel better, this is not the case with elderly people who cannot communicate their feeling of thirst and for people with impaired thirst drive, a condition known as adipsia
We propose modeling the dehydration problem as a regression problem to predict the last drinking time using features extracted and constructed from multiple wearable sensors (accelerometer, magnetometer, gyroscope, galvanic skin response (GSR) sensor, photoplethysmography (PPG) sensor, temperature, and barometric pressure sensor); For this purpose, we recorded a total of 3386 min for these sensors for 11 subjects under fasting and non-fasting conditions; Sensors 2022, 22, 1887

Summary

Introduction

Following the tremendous technological advances in the design of systems on a chip (SoCs), the development and use of wearable devices have remarkably achieved high growth rates in the last few years. Wearable devices for different biosignals’ acquisition have been researched [1], and challenges to and opportunities for wearable device design, biosignal analysis, and interpretation have been discussed [2] These advances in the research and industry of wearable devices drive the research and development efforts to use wearable devices to monitor, analyze, and learn from different sensors’ data for different people’s healthcare applications. People whose work requires laborious non-stop effort in very hot or very cold weather, such as soldiers and construction workers, may miss drinking regularly. These cases have great need for dehydration monitoring. There are relatively few studies that have worked in this area, which we briefly review for the used features, techniques, and evaluation metrics

Results

Discussion

Conclusion