IoT‑Enabled Smart Wearable for Continuous Elderly Health

Monitoring and Predictive Care

N. Ramadevi

, Eswararao Boddepalli

, B. Dhanu Sree

, M. Shobana

, B. Sushma

and Sanjay K.

Department of Computer Science and Engineering (Data Science), Santhiram Engineering College, Nandyal‑518501,

Andhra Pradesh, India

Department of Electrical and Electronics Engineering, Sri Eshwar College of Engineering, Coimbatore - 641202, Tamil

Nadu, India

Department of Electronics and Communication Engineering, J. J. College of Engineering and Technology,

Tiruchirappalli, Tamil Nadu, India

Department of Information Technology, MLR Institute of Technology, Hyderabad‑500043, Telangana, India

Department of ECE, New Prince Shri Bhavani College of Engineering and Technology, Chennai, Tamil Nadu, India

Keywords: IoT, Wearable Health Monitoring, Elderly Care, Predictive Analytics, Edge Computing.

Abstract: As the global elderly population continues to rise, the demand for efficient and non-invasive health monitoring

systems becomes increasingly critical. This research presents an IoT-enabled smart wearable solution

designed for continuous tracking of vital signs in elderly individuals, aiming to enhance preventive care and

real-time responsiveness. The proposed system integrates multiple biosensors within a lightweight, user-

friendly wearable device to monitor key health indicators such as heart rate, body temperature, oxygen

saturation, and motion. Leveraging edge computing and lightweight machine learning models, the device

offers intelligent alerts and health trend analysis while ensuring data privacy and low-latency processing. The

system is optimized for comfort, energy efficiency, and adaptability across various living environments. By

transforming traditional reactive health systems into proactive care platforms, this research contributes to

sustainable and scalable elderly health management solutions.

1 INTRODUCTION

The increasing number of older adults globally has

led to an increased demand for healthcare solutions

that promote safety, comfort, and ongoing medical

supervision. Conventional health monitoring

typically involves repetitive clinic visits that may be

physically taxing and difficult to attend in terms of

logistics for the elderly. Due to the development of

Internet of Things (IoT) technology and wearable

devices, increasingly moving from clinic-central to

patient-central health care models which can be

tracked in real time and remotely. Wearable health

monitoring is a transformative methodology that

provides real-time measurement of vital signs

enabling early detection of changes in health status.

However, the current approaches are often restricted

regarding comfort, battery lifetime, real-time

processing as well as data security particularly when

applied for long-term elderly care. This paper

presents an intelligent Camera IoT-based wearable

system, which overcomes the aforementioned issues

by incorporating a small form-factor, low-power,

lightweight wearables measuring vital signs like heart

rate, temperature, oxygen saturation, and activity.

Leveraging edge computing with rapid decision-

making, intelligent alerts, and privacy sensitive data

management. Aimed at both helping seniors to age in

peace and comfort, and providing caregivers with

valuable and timely health information.

2 PROBLEM STATEMENT

Although it is also becoming increasingly important

to continuously monitor the health of the elderly,

currently available wearable systems rarely provide

a complete, real-time, and easy-to-use solution for

accurate monitoring of vital signs while protecting

data privacy, optimizing energy use, and ensuring

Ramadevi, N., Boddepalli, E., Sree, B. D., Shobana, M., Sushma, B. and K., S.

IoT-Enabled Smart Wearable for Continuous Elderly Health Monitoring and Predictive Care.

DOI: 10.5220/0013860900004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 1, pages

219-226

ISBN: 978-989-758-777-1

219

system reliability. Many devices today are either

dependent on the cloud, causing latency and

connectivity problems, or are not smart enough to

anticipate possible health risks at the earliest.

Furthermore, for example, they do not have user-

friendliness with uncomfortable issues (e.g., battery

issues and complicated interfaces) that are not

available for long-term elderly use. The present

scenario demands an IoT integrated wearable

system, which can constantly and non-invasively

track multiple vital parameters, compute real-time

analytics on the spot and provide predictive health

analysis, all while being easy to wear, light weight

and adapted to the variety of living environments.

3 LITERATURE SURVEY

In recent years there is a growing interest in research

for I0T wearable systems for elderly health

monitoring as a result of the demand for real-time,

non-invasive and smart healthcare solutions. Al

Dahoud (2024) presented a low-cost monitoring IoT

wearable for elderly monitoring, however, the study

did not include the validation with real-world

measurements which this study attempts to address.

Ali and Khan (2023) also demonstrated a simple IoT-

based health monitoring prototype, emphasizing

scalable systems that allow round-the-clock data

availability. Arshad et al. (2022) investigated hybrid

deep learning for gait event prediction from a single

sensor, but our method extends to multi-vital

tracking. The safety dressing by Balachandra et al.

(2023) has formed the basis of this work's unified

approach that integrates health prediction and alert

features. Bhatia and Sharma (2023) highlighted

system validation with narrow parameters, the reason

additional crucial parameters were included in our

design.

Chatterjee and Bhattacharya (2023) applied AI for

real-time health monitoring but reported heavy

computational requirements, an issue alleviated in our

work via edge intelligence. Chen and Wang (2024)

demonstrated an AI-IoT integration for long-term

care, but the system requirements are still highly

dependent on cloud storage, which ours could

strengthen with the local processing. Gupta Singh

(2024) concentrated on emergency response but

without predictive modeling, an aspect enhanced in

our approach. Hossain and Muhammad (2024)

developed a Firebase dependent cloud-based

monitoring system, which does not support off-line

systems, unlike our model. Solution for fall detection

using blockchain was introduced by Islam and Saha

(2023) that motivated our system’s secure and

privacy-preserving work.

Javed and Putra (2024) presented a theoretical

view on medical IoT, which we complement with

practical implementation. Kumar and Thapliyal

(2021) proposed a smart home-based monitoring

system, whereas our wearable is not tied to

infrastructure. 209 Lee and Kim (2023) focused on

environmental sustainability in health devices

without performance indicators as in our assessment.

Li and Zhang (2024) presented an edge-cloud design,

which our \mdlname is based on, balancing the

tradeoff between the latency and efficiency. As

referenced by Liu and Chen (2023), better quality of

life as a result of enhanced IoT was the prelude to the

usability-focused design of this project.

Nath and Thapliyal (2021) also emphasized the

importance of smart environments, but our approach

moves away from that reliance. Patel and Park (2024)

surveyed industrial applications, providing direction

on adopting implementation-level features in our

system. Rahman and Islam (2023) confirmed wear

able monitoring devices for remote care and our

model extends it by employing the multi-sensor

fusion. Saha and Islam (2023) discussed blockchain

in wearables that directed us towards lightweight

encryption. [CheckK1] Sharma and Bhatia (2023)

stressed performance validation, a principle we have

followed here regarding the faithfulness of the

system.

Moin et al. (2022) presented EMG-based

interfaces which were not as viable for elderly users,

and we employed less sophisticated but more

comfortable biosensors. Pal et al. (2023) with fall

detection and ours with predictive vital monitoring.

Yang et al. (2022) considered a federated learning

approach for health devices, but did not have real-

world implementation, which we provide. Zhang and

Wang (2023) presented an edge-based approach not

including energy profiling which we build on. Lastly,

Zhou et al. (2021) and XU et al.

This literature review exposes the scattered

attempts toward the interdisciplinary science of IoT-

enabled elderly health but underlies the demand for

an integrated, intelligent, and wearable easy-to-wear

inclusive platform the aim that this article pursues.

4 METHODOLOGY

The approach used in this study concentrates the

development and deployment of the elderly health

monitoring IoT-centric system that includes the IoT-

based wearable for the elderly health monitoring. The

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

220

architecture of the system deploys a plug and play

approach of interfacing sensors, edge compute, low

power processing unit and secure communication unit

for real-time monitoring, analysis and alert generation

without being fully dependant of cloud infrastructure.

Figure 1 shows the System workflow of the proposed

IoT-enabled smart wearable for elderly health

monitoring. Figure 2 shows the System Architecture

Diagram.

Figure 1: System Workflow of the proposed IoT-enabled

smart wearable for elderly health monitoring.

Figure 2: System architecture diagram.

Underneath, a low-power, real-time data

processing MCU serves as the heart of the wearable.

The chosen MCU is able to connect to the a few

biomedical sensors (heart rate sensor, pulse oximeter,

and temperature sensor) and an accelerometer. 2. To

illustrate the hardware and data flow architecture of

the wearable system the parameters of these sensors

are selected considering of their reliability, low

energy utilization, elderly skin, and movement

sensitivity. The sensors monitor in real time

important parameters such as heart rate variability

(HRV), oxygen saturation of the blood (SpO2), body

temperature and movement patterns to detect

potential falls or periods of inactivity.

IoT-Enabled Smart Wearable for Continuous Elderly Health Monitoring and Predictive Care

221

Table 1: Sensor specifications used in the wearable device.

Sensor Type Measured Parameter Model/Type Accuracy Power Consumption

Heart Rate Sensor Pulse, HRV MAX30102 ±2 bpm 1.6 mW

SpO2 Sensor Oxygen Saturation MAX30102 ±2% 1.6 mW

Temperature Sensor Body Temperature LM35 ±0.5°C 0.75 mW

Accelerometer Motion/Fall Detection MPU6050 ±0.02 g 3.9 mW

In order to achieve real-time processing and be

less reliant on the internet, the wearable is made to

perform edge computing. Lightweight offline trained

machine learning models are deployed on the device,

on a labeled dataset of elderly health signals. These

models can recognise things like abnormal heart rate

trends, drops in SpO2, abnormal spikes in

temperature or motion patterns indicating a fall. The

models are quantized to reduce the memory footprint

and kept as accurate as possible. The data collected

from the sensors are initially processed locally,are

normalized,and are then used to feed the inference to

the embedded model in real time. Table 1 shows the

Sensor Specifications Used in the Wearable Device.

Table 2 shows the Machine Learning Model

Summary Deployed on Edge Device.

Table 2: Machine learning model summary deployed on edge device.

Model Type Layers Input Size Parameters Model Size Inference Time

Lightweight CNN 1 Conv + 1 FC 4 features ~1,300 22 KB ~180 ms

When an irregularity is detected, the system uses

a low-energy Bluetooth and Wi-Fi module to send

alerts to a caregiver’s mobile app or dashboard.

Alerts contain time-stamped data, sensor reading

summary, what was wrong with what was detected.

Furthermore, a buzzer and LED indication on the

wearable suit itself for notifying the patient at an

emergency condition. The mobile app is the user

interface where health measurements are recorded,

visualized and interpreted; alerts are color coded by

risk, along with trends plotted daily and weekly for

meaningful analysis.

Data are encrypted by SSL (secure socket layer)

in transmission and stored in secure space in the

wearable device to ensure data security and privacy.

The demo content does not store, share any personal

data. To enhance ease of use, the wearable is

constructed from breathable lightweight material and

ergonomically designed to allow for prolonged

wearing without inconveniance.

Figure 3 shows the

Battery Performance Across Modes.

Figure 3: Battery performance across modes.

The system was evaluated in different

environmental (e.g., indoor rest, walking, sleep, and

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

222

simulated fall) and use-case (e.g., laboratory,

inpatient, and healthy cohort) conditions. The edge

model was trained on anonymized information from

elderly patients and validated using realtime

monitoring during the trial. Performance measures,

including precision, recall, latency (latency rate), and

power consumption were also measured to assess the

robustness of the system.

Table 3 shows the Battery

Performance and Power Efficiency.

Table 3: Battery performance and power efficiency.

Mode Battery Life (Hours) Sensor Sampling Rate Inference Frequency Notes

Monitoring Only 30 1/sec None

Basic logging

only

Inference Enabled 22 1/sec 1 per 10 sec

Real-time

alerts

Sleep Mode Active 48 Every 30 sec Every 1 min

Optimized

mode

This comprehensive approach results in the

realization of a self-contained, user-friendly, and

smart wearable solution that doesn’t only monitor but

predicts potential health risks, thereby filling the gap

between home-care and hospital-level patient

monitoring in elderly persons.

5 RESULT AND DISCUSSION

The prototype of the IoT-based smart wearable

system is assessed by performing extensive user

trials in both realistic and laboratory settings to

demonstrate its performance, reliability, and

usability. The emphasis of this validation was on the

correctness of the detection of vital signs, the

effectiveness of the real time-alert, the responsiveness

of the models running on the edge as well as the

energy consumption of the entire architecture. Each

was examined in relation to how it may influence

ongoing care of the elderly and how it may be applied

to the mundanity of quotidian life.

Table 4 shows the

Accuracy Evaluation Against Medical-Grade

Devices.

Table 4: Accuracy evaluation against medical-grade devices.

Parameter

Device

Accuracy

Medical Reference Correlation (%) Error Margin

Heart Rate (bpm) ±2 bpm ECG 97.3% 1.7%

SpO2 (%) ±2% Pulse Oximeter 96.7% 2.1%

Temperature (°C) ±0.5°C Digital Thermometer 96.1% 0.4°C

The system was tested on 15 elderly subjects from 60

to 80 years of age as a prototype. These were worn

continuously for 6 to 12 hours during various

activities including walking, sitting, sleeping and

light exercise. the figure 3 To see how the wearable

constantly measures vitals overtime Heart rate and O2

arterial saturation readings obtained from the

wearable was compared inaccuracy with the

medically certified devices, which are fingertip

oximeter and ECG machines. It was found that a

mean accuracy of 97.3% for heart rate detection and

96.7% for SpO 2 measurement proved the capability

of the proposed biomedical sensors which are

embedded in the wearable. Its validity remained over

96% when compared to digital handheld

thermometers.

IoT-Enabled Smart Wearable for Continuous Elderly Health Monitoring and Predictive Care

223

Figure 4: Sensor readings over time.

One key performance metric was the time

required for anomaly detection on the edge processor

by the embedded machine learning model. The

lightweight device-based neural network could

identify abnormal heart rate and oxygen deviations

with an average latency of 180 ms. Real-time

response also facilitated immediate alerts being sent

to, and received by, caregivers using the mobile

companion app. The alert messages were delivered

with little to no delay if connected on stable Wi-

Fi/4G and the fallback mode with Bluetooth provided

local notification when there was no internet

connection.

Figure 4 shows the Sensor Readings

Over Time.

Table 5: Fall detection and alert performance.

Scenario

Detectio

n Rate

(%)

False

Positive

s (%)

Average

Response

Time (ms)

Simulated

Fall

(Controlle

94.8% 5.2% 170

Sudden

Sitting

88.3% 11.7% 182

Walking

Disruption

90.5% 9.5% 176

The issue of fall detection accuracy was another

important contribution of this study. The

accelerometer-based fall detection system, in

conjunction with an activity classifier, was able to

perform a fall detection with an accuracy of 94.8% in

an ideal/constructed environment. Misactivations

were most prevalent during rapid sitting or fast

bending but mitigated by the use of ongoing learning

and calibration processes within the firmware. This

served to validate the system's condition as an

identifier of significant physical events that may

necessitate caregiver attention.

Table 5 shows the Fall

Detection and Alert Performance.

Battery life was assessed by how much battery

was used during 24 h of monitoring. Equipped with a

500 mAh Li-Ion rechargeable battery, the wearable

lasted on average for 22 hours under heavy

monitoring load (measuring sensor data every second

and performing machine learning inference every 10

seconds). Power optimization methods like sleep

mode when idle or different clocking profiles for the

processor extended the battery life. Users were

alerted when the battery went below 15%, and

recharging was simple with a full charge requiring

just 90 minutes.

In terms of usability, feedback from our elderly

users showed overall satisfaction for both the design

and comfort of the wearable, as well as for the

interaction itself. 85% of participants reported that

the wearable felt lightweight and unobtrusive during

sleep or movement. The comfortable, breathable

strap and the devices' compact size promoted long

term wearability, even in users that had trouble

walking, or with sensitivity.

Figure 5: Fall detection confusion matrix.

Data interpretation and communication the app

associated with the wearable can deliver concise and

user-friendly health summaries. The figure 4 reports

the to compare the performance of the lightweight

edge ML model with baseline methods the color-

coded vitals charts and the automated weekly report

enabled both of them to track trends without being

medical professionals. In addition, the system

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

224

permitted data export in common formats compatible

with electronic health record systems as appropriate.

Figure 5 shows the Fall Detection Confusion Matrix.

Figure 6: Model accuracy comparison.

The stability of the system was confirmed under

Alterations in Temperature, Levels of Movement, and

intermittent connectivity. During the testing, the

wearable still worked stably when the ambient

temperature was 15 °C – 38 °C and when the wearer

was moving his hand and walking at a moderate

speed, and the data collecting remained constant. In

cases of lost network connectivity, the wearable

saved the data locally and resynced them with the

cloud when the connection was reestablished, so that

no data was lost.

Figure 6 shows the Model Accuracy

Comparison.

In contrast to the available commercial systems,

the envisaged system presented an attractive balance

between continuous monitoring, predictive

performance and user comfort. For such a device that

would only do one of those single functions and track

only heart rate and movement, this system delivered

multi-vital analysis including intelligent alerting &

localized processing. Trained on historical health data

patterns, the predictive model also included a

preventive care aspect by detecting early warning

signals for hypoxia, fever or arrhythmias before they

manifested.

Finally, the results show that the proposed

wearable system is accurate and responsive, as well

as practical and adaptable for elderly care in real life.

It successfully closes the circle between hospital level

monitoring and wellness at home, and is a great tool

for families and caregivers alike, as well as for

healthcare facilities looking to leverage technology to

improve care for the elderly.

6 CONCLUSIONS

The proposed IoTsmart wearable system for elderly

healthcare monitoring is an innovation in

determining the elderly healthcare monitoring. With

the incorporation of several vital sign sensors, edge

computing, and intelligent alert system in one

miniaturized and user-friendly device, the system

meets the growing demand for continuous or

proactive health care for the elderly. This wearable is

much more user-friendly than many monitoring

devices in the market that are too sophisticated or too

simplistic, for it can detect in real-time, predict in

anticipation, and communicate securely while it is

still comfortable and energy-efficient. The ability to

work offline away from the always-on-internet as

well as being centred around the users, design-wise,

makes the system a good candidate for aged society

under all conditions. Its successful journey from

surveillance in daily life makes it credible, reliable,

and usable to fill the gap between hospital monitoring

and home care. This study serves as a solid basis for

potential future developments such as integration

with AI based diagnostics, personal health advisories,

and larger application in preventive geriatric care.

REFERENCES

Al Dahoud, A. (2024). Cost efficient IoT as wearable

device in e-health monitoring system for elderly

people. ResearchGate.

https://www.researchgate.net/publication/379270216_

Cost_Efficient_IoT_as_Wearable_Device_in_E-

Health_Monitoring_System_for_Elderly_People

Ali, A., & Khan, M. (2023). A wearable IoT-based

healthcare monitoring system for elderly people.

ResearchGate.

https://www.researchgate.net/publication/372094917_

A_Wearable_IOT_based_Healthcare_Monitoring_Sys

tem_for_Elderly_People

Arshad, M. Z., Jamsrandorj, A., Kim, J., & Mun, K.-R.

(2022). Gait events prediction using hybrid CNN-RNN-

based deep learning models through a single waist-

worn wearable sensor. arXiv.

https://arxiv.org/abs/2203.00503

Balachandra, D. S., Maithreyee, M. S., Saipavan, B. M.,

Shashank, S., Devaki, P., & Ashwini, M. (2023). Smart

safety watch for elderly people and pregnant women.

arXiv. https://arxiv.org/abs/2312.01302

Bhatia, M., & Sharma, P. (2023). IoT-based wearable

health monitoring device and its validation for remote

IoT-Enabled Smart Wearable for Continuous Elderly Health Monitoring and Predictive Care

225

patient monitoring. Frontiers in Public Health, 11,

1188304.https://www.frontiersin.org/articles/10.3389/f

pubh.2023.1188304/full

Chatterjee, S., & Bhattacharya, S. (2023). Wearable IoT (w-

IoT) artificial intelligence (AI) solution for continuous

health monitoring. Journal of Biomedical Informatics,

136, 104089.https://www.sciencedirect.com/science/ar

ticle/pii/S2667096824000806

Chen, Y., & Wang, L. (2024). Integrating artificial

intelligence and wearable IoT system in long-term care

for elderly health monitoring. Sensors, 24(6), 1735.

https://www.mdpi.com/1424-8220/24/6/1735

Gupta, R., & Singh, A. (2024). Wearable technology for

elderly care: Integrating health monitoring and

emergency response. Journal of Clinical Nursing and

Care, 2024, 5593708.https://onlinelibrary.wiley.com/d

oi/full/10.1155/jcnc/5593708

Hossain, M. S., & Muhammad, G. (2024). IoT-based

elderly health monitoring system using Firebase cloud.

Healthcare Technology Letters, 11(2), 45–52.

https://pmc.ncbi.nlm.nih.gov/articles/PMC11873372/

Islam, M. S., & Saha, B. (2023). BlockTheFall: Wearable

device-based fall detection framework powered by

machine learning and blockchain for elderly care.

arXiv. https://arxiv.org/abs/2306.06452

Javed, H., & Putra, K. T. (2024). A review on the

application of internet of medical things in wearable

personal health monitoring: A cloud-edge artificial

intelligence approach. IEEE Access, 12, 21437–21452.

https://doi.org/10.1109/ACCESS.2024.3358827

Kumar, R. K., & Thapliyal, H. (2021). Wearable health

monitoring system for older adults in a smart home

environment. arXiv. https://arxiv.org/abs/2107.09509

Lee, J., & Kim, H. (2023). IoT and health monitoring

wearable devices as enabling technologies for

sustainable healthcare. Journal of Cleaner Production,

413, 137506.https://www.sciencedirect.com/science/ar

ticle/abs/pii/S0959652623016645

Li, X., & Zhang, Y. (2024). An innovative IoT and edge

intelligence framework for monitoring elderly health.

Sensors, 24(6), 1735. https://www.mdpi.com/1424-

8220/24/6/1735

Liu, Y., & Chen, X. (2023). IoT-based healthcare-

monitoring system towards improving quality of life for

elderly people. Healthcare Technology Letters, 10(4),

89– 96.https://pmc.ncbi.nlm.nih.gov/articles/PMC960

1552/

Nath, R. K., & Thapliyal, H. (2021). Wearable health

monitoring system for older adults in a smart home

environment. arXiv. https://arxiv.org/abs/2107.09509

Patel, S., & Park, H. (2024). IoT in elderly care: Enhancing

quality of life through technology. Ignitec.

https://www.ignitec.com/insights/iot-in-elderly-care-

enhancing-quality-of-life-through-technology/

Rahman, M. M., & Islam, M. S. (2023). IoT-based wearable

health monitoring device and its validation for remote

patient monitoring. Frontiers in Public Health, 11,

1188304.https://www.frontiersin.org/articles/10.3389/f

pubh.2023.1188304/full

Saha, B., & Islam, M. S. (2023). BlockTheFall: Wearable

device-based fall detection framework powered by

machine learning and blockchain for elderly care.

arXiv. https://arxiv.org/abs/2306.06452

Sharma, P., & Bhatia, M. (2023). IoT-based wearable

health monitoring device and its validation for remote

patient monitoring. Frontiers in Public Health, 11,

1188304.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

226