| Journal of Information and Communications Technology:

Algorithms, Systems and Applications

Received: 10 May 2026; Revised: 12 June 2026; Accepted: 22 June 2026; Published Online: 25 June 2026.

J. Inf. Commun. Technol. Algorithms Syst. Appl., 2026, 2(2), 26308 | Volume 2 Issue 2 (June 2026) | DOI: https://doi.org/10.64189/ict.26308

This article is licensed under Creative Commons Attribution NonCommercial 4.0 International (CC-BY-NC 4.0)

SeizureSense: A Wearable Epilepsy Monitoring

System

Dnyaneshwar S. Mantri,* Sushilkumar S. Salve, Yogesh Budhalkar,* Sakshi Bhaltilak and Dnyandev More

Department of Electronics & Telecommunication Engineering, Sinhgad Institute of Technology, Lonavala Savitribai Phule Pune

University, Pune, Maharashtra, 410101, India

*Email: dsmantri.sit@sinhgad.edu (Dnyaneshwar S. Mantri), budhalkaryogesh@gmail.com (Yogesh Budhalkar)

Abstract

Epilepsy is one of the most common chronic neurological disorders, affecting approximately 10–12 million

people in India and nearly 65 million people worldwide. It is considered the fourth most common neurological

condition after migraine, Alzheimer’s disease, and stroke. Epileptic seizures can cause serious health risks such

as falls, head injuries, fractures, burns, breathing difficulties, and convulsive status epilepticus, creating anxiety

among patients, parents, and caregivers. To address this problem, this paper proposes a wearable IoT-based

seizure alert system for real-time seizure detection and emergency notification. The proposed system uses an

ESP32-C3 microcontroller, MAX30102 heart rate and SpO₂ sensor, MPU6050 accelerometer and gyroscope, and

an EMG sensor to monitor physiological parameters and abnormal muscle movements associated with seizures.

When a seizure is detected, the device automatically sends an SMS alert with the user’s location to caregivers or

family members and activates a speaker alert to notify nearby people. Prototype-level testing under controlled

laboratory conditions demonstrated an estimated seizure detection accuracy of 92.5% of 5.2%, and a response

time of 3–5 seconds, demonstrating that it is a reliable, real-time, and efficient solution for epilepsy monitoring.

Keywords: Epilepsy monitoring; ESP32-C3; Seizure detection; Wearable healthcare; Internet of Things (IoT); Multi-

sensor fusion; Heart rate and SPO

Monitoring; Emergency alert system.

1. Introduction

Epilepsy is a serious neurological disorder affecting millions of individuals worldwide.

[1]

It is characterized by

sudden and unpredictable seizures, resulting from abnormal electrical activity in the brain.

[2]

These seizures can

vary in intensity, causing loss of consciousness, involuntary movements, and serious physical injuries.

[3]

The

unpredictability of seizures poses severe risks, such as falls, accidents, and even sudden unexpected death in

epilepsy (SUDEP).

[4]

Consequently, real-time seizure detection and alert systems are critical for ensuring timely

intervention and enhancing patient safety.

[5-7]

In the medical field, traditional methods for epilepsy monitoring

include video-electroencephalography (v-EEG) and electrocardiography (ECG).

[8-10]

These methods are

commonly used in hospitals to diagnose and analyze seizure patterns by capturing brain activity, body

movements, and heart rate fluctuations.

[11]

However, these techniques have several limitations, including high

costs, the need for specialized clinical settings, and prolonged monitoring, making them impractical for

everyday use.

[12]

Due to these limitations, there has been a growing interest in wearable IoT-based seizure

detection systems that offer continuous real-time monitoring while allowing patient mobility.

[13-18]

With

advancements in embedded systems and sensor technology, researchers have developed motion-based seizure

detection using accelerometers and gyroscopes, such as MPU6050, to identify abnormal movement patterns.

[19-

22]

Additionally, heart rate sensors are increasingly used to monitor sudden physiological changes associated

with seizures.

[23-26]

Several studies have explored microcontroller-based solutions, leveraging devices like

ESP32 or Arduino to process sensor data and trigger emergency alerts.

[27-29]

Furthermore, GSM modules

(SIM800L/SIM900A) have been integrated to send automated SMS notifications to caregivers, ensuring

immediate assistance, even in remote areas without Wi-Fi connectivity.

[30,31]

Our proposed project builds upon

these existing technologies by introducing a wearable epileptic seizure alert system utilizing ESP32C3 XIAO as

the central processing unit. The system integrates an MPU6050 motion sensor, a heart rate sensor, and a DF-

Player Mini with a speaker module to detect seizure activity and provide both audible and SMS-based alerts. By

incorporating GPS tracking, the system enables caregivers to quickly locate the patient in case of an

emergency.

[32,33]

Unlike hospital-based monitoring methods, our lightweight, portable, and cost-effective

solution offers continuous real-time monitoring, improving response times and overall patient safety.

Addressing key challenges such as power efficiency, false alarms, and latency in alert transmission, our research

contributes to the advancement of wearable healthcare technology for epileptic seizure detection and

emergency response.

[29,34,35]

The system activates an emergency response mechanism, which includes: Audible

Alert System – A DF-Player Mini module with a speaker generates a warning sound to alert nearby individuals.

Emergency SMS Notification – A GSM module (SIM800L/SIM900A) automatically sends an alert message to

caregivers, ensuring quick intervention. GPS Tracking for Emergency Location – A GPS module tracks the

patient’s location and shares it with caregivers in real time.

Compared to traditional hospital-based EEG monitoring, our system is designed to be lightweight, cost-

effective, and user- friendly, allowing patients to move freely while benefiting from continuous real-time

monitoring. This approach significantly reduces response time during seizures, making it an effective and

practical solution for epilepsy management. The primary contributions of this work include: Development of a

wearable, real-time epileptic seizure detection system with minimal intrusion into the patient’s daily life.

Integration of multi-sensor technology (motion and heart rate sensors) for accurate seizure detection.

Implementation of an emergency alert system combining audible alarms, SMS notifications, and GPS tracking

for immediate assistance. Optimization of power consumption to ensure prolonged battery life for

uninterrupted monitoring.

2. Literature survey

Epilepsy is a chronic neurological disorder characterized by recurrent and unpredictable seizures that pose

serious risks to patient health and safety.

[1,2]

Over the years, various seizure detection and alert systems have

been developed using physiological signals, motion sensors, and wireless communication technologies.

[3,4]

Existing seizure detection approaches can be broadly classified into EEG-based systems, motion-based systems,

physiological signal-based systems, and wearable IoT-based healthcare systems.

[5,14,24]

Early research by Shoeb

and Guttag demonstrated the use of electroencephalogram (EEG) signals combined with machine learning

techniques for seizure detection.

[7]

Although EEG-based systems provide high detection accuracy, they require

complex signal processing, clinical-grade equipment, and continuous monitoring, making them less suitable for

portable wearable applications.

[7,9,10,36,37]

To improve wearability and real-time monitoring, researchers explored

motion-based seizure detection using accelerometers and gyroscopes. Motion-based systems effectively detect

repetitive and high-amplitude body movements associated with seizure activity, but they often generate false

positives during normal physical activities such as running or sudden movements.

[4,5,21,30]

Physiological signal

monitoring has also gained considerable attention for seizure detection. Van Elmpt et al. reported significant

heart rate variations during seizure episodes, while Poh et al. demonstrated that physiological indicators such

as heart rate variability improve seizure detection reliability.

[12,13]

These approaches help identify abnormal

physiological changes that occur during epileptic events and enhance detection accuracy when combined with

motion analysis.

[32,38]

Recent advancements in wearable healthcare technology and IoT systems have enabled the development of

real-time remote seizure monitoring solutions. Wearable sensor-based healthcare systems provide continuous

physiological monitoring and improve patient mobility compared to traditional hospital-based monitoring

methods

.[5,15-18]

In addition, GSM-based emergency alert systems have been integrated into wearable healthcare

devices to send automatic notifications to caregivers during emergencies, ensuring timely medical

assistance.

[8,31]

Machine learning, deep learning, multimodal sensing, and adaptive IoT architectures have

further enhanced the reliability of seizure detection systems.

[22,35-39]

Recent studies have also demonstrated the

benefits of wearable digital health technologies, energy-efficient embedded systems, and clinical seizure

detection devices for long-term epilepsy monitoring.

[7,15,26,29,40-42]

These studies indicate that combining motion

sensing, physiological monitoring, machine learning, and wireless communication technologies can

significantly improve the reliability and effectiveness of wearable epileptic seizure detection systems.

2.1 Research gap

Despite significant advancements, existing systems suffer from limitations such as high false alarm rates,

incomplete physiological monitoring, and limited real-time alert capabilities. Most solutions rely on either

motion-based or physiological parameters independently, which fails to capture the complete characteristics

of seizure events.

2.2 Proposed-approach

To address these limitations, the proposed system integrates motion, heart rate, SpO₂, and EMG signals within

a wearable IoT platform. By employing a multi-sensor fusion approach along with real-time alert mechanisms

and GPS-based location tracking, the system improves detection accuracy, reduces false positives, and ensures

timely emergency response. Table 1 presents a comparison of existing seizure detection systems based on their

sensing techniques, features, and limitations. It highlights that most existing approaches lack comprehensive

multi-sensor integration and reliable real-time alert mechanisms.

Table 1: Comparison of existing seizure detection system.

Ref No.

Author/ Year

Methodology used

Technologies

Key

features

Results/Outcome

Limitations

[41]

Djemal et al.

(2023)

EMG-based

seizure

classification

Wearable EMG

Sensor, ML

Automatic

seizure

classification

using muscle

activity

High seizure

classification

performance

Limited to EMG

signals only

[38]

ElSayed et al.

(2023)

IoT-based

adaptive seizure

detection

IoT, Machine

Learning

Adaptive

architecture for

epilepsy

monitoring

Improved remote

healthcare

monitoring

Requires ML

training and

optimization

[35]

Ingolfsson et

al. (2023)

Energy-efficient

embedded

detection

Embedded AI,

Wearable Sensors

Low-power

seizure detection

system

Reduced

computational

power consumption

Limited clinical

validation

[15]

Donner et al.

(2024)

Wearable digital

health monitoring

Wearable Sensors,

Digital Health

Platform

Continuous

epilepsy

monitoring

Improved patient

management

Commercial

deployment

challenges

[16]

Bernini et al.

(2024)

Ambulatory

seizure detection

Wearable Devices

Real-world

seizure detection

outside hospitals

Enhanced

ambulatory

monitoring

Variable

performance across

seizure types

[17]

Sasseville et

al. (2024)

Wearable seizure

detection review

Wearable

Biosensors

Analysis of

device

performance and

usability

Demonstrated

practical clinical

benefits

Device-specific

limitations

[26]

Baumgartner

et al. (2025)

Seizure detection

device review

Wearable Detection

Devices

Clinical

assessment of

seizure detectors

Comprehensive

evaluation of

technologies

Device

performance varies

[22]

Faust et al.

(2025)

Machine learning

wearable

detection

Wrist-worn

Sensors, ML

Detection of

multiple seizure

types

Improved wearable

detection capability

Dependent on

training data

quality

[43]

Bhagubai et

al. (2025)

Multimodal

epilepsy dataset

EEG, ECG, EMG,

Accelerometer,

Gyroscope

Large wearable

epilepsy dataset

Supports

development of

advanced

algorithms

Dataset-focused

study

[19]

Jeppesen et

al. (2025)

Clinical wearable

ECG detection

Wearable ECG

Device

Automated

seizure detection

Clinical validation of

wearable

monitoring

Limited to ECG-

based detection

3. Real-time methodology

In this section, we present a wearable IoT-based epileptic seizure detection and alert system designed for real-

time monitoring and emergency response. The system aims to provide an efficient, non-intrusive, and reliable

solution for individuals with epilepsy, offering continuous health monitoring and instant alert mechanisms. The

overall flow of our proposed system is shown in Fig. 1. Flowchart of Seizure Alert System, consisting of two

main phases: signal acquisition and processing and emergency alert mechanism, is shown in Fig. 2. Each of these

steps is thoroughly explained in the following subsections. The effectiveness of seizure detection depends on

accurate data collection and efficient signal processing. The proposed wearable system utilizes multiple

biosensors embedded within the device, including an MPU6050 motion sensor, a pulse sensor for heart rate

monitoring, and a GPS module, all controlled by the ESP32C3 XIAO microcontroller. These sensors continuously

collect real-time physiological and motion-related data from the user, while the ESP32C3 processes the

incoming sensor data to identify abnormal conditions associated with epileptic seizures. The MPU6050 sensor,

consisting of a 3-axis accelerometer and gyroscope, continuously monitors body movements and orientation.

Seizure activity is characterized by abnormal movement patterns such as rhythmic shaking, sudden jerks, and

uncoordinated body movements. To differentiate seizure activity from normal daily movements, a threshold-

based motion detection algorithm is implemented. The algorithm analyzes motion features including

acceleration spikes beyond normal movement thresholds, angular velocity variations caused by involuntary

jerking, and the frequency and duration of repetitive body movements. An adaptive filtering technique is

applied to remove unwanted noise from the motion data while preserving critical seizure- related features. If

the detected motion parameters exceed predefined threshold values for a sustained period, the system

identifies the condition as a possible seizure event. In addition to motion analysis, the wearable device

continuously monitors heart rate variations using a pulse sensor.

Fig. 1: Block Diagram of Seizure Alert System

Fig. 2: Flow diagram of seizure alert system.

Epileptic seizures often cause physiological changes such as tachycardia (rapid heart rate) or bradycardia (slow

heart rate). The ESP32C3 microcontroller records and analyzes baseline heart rate patterns, sudden heart rate

surges or drops, and heart rate variability during seizure events. Moving average filtering and peak detection

algorithms are employed to improve measurement accuracy and reduce signal noise. The system cross-

validates abnormal motion data with heart rate fluctuations to minimize false positives, ensuring that normal

physical activities such as running or sudden movements do not trigger unnecessary seizure alerts. The

ESP32C3 XIAO serves as the central processing unit responsible for real-time seizure detection and decision-

making using a decision tree-based anomaly detection algorithm. The algorithm compares detected motion

patterns and physiological changes with predefined seizure markers to determine whether the user is

experiencing a seizure.

When both abnormal movement and irregular heart rate patterns are detected simultaneously, the system

confirms the seizure event and activates the emergency alert mechanism. To reduce computational load and

improve battery efficiency, the system operates in a low-power mode whenever no abnormal activity is

detected. Once a seizure is confirmed, the system automatically activates emergency alert features to notify

caregivers, family members, and nearby individuals. A DF-Player Mini module connected to a mini speaker

plays a pre-recorded emergency voice message to alert nearby people and request immediate assistance. This

audible alert mechanism is particularly useful when the affected individual is alone or in public places where

rapid help may be required. Table 2 lists the predefined threshold values used for seizure detection. These

thresholds are selected based on observed physiological and motion characteristics during tonic– clonic seizure

events. A seizure is detected only when multiple parameters simultaneously exceed their respective thresholds,

thereby reducing false alerts. To help emergency responders locate the affected person, a GPS module

continuously updates the user’s location. The longitude and latitude coordinates are transmitted via SMS to

caregivers, enabling them to reach the individual quickly. The location data is refreshed every few seconds,

ensuring accurate real-time tracking until the seizure episode ends. This feature is particularly crucial for

individuals who experience seizures in public places, while traveling, or living alone.

Table 2: Threshold values used for seizure detection.

Parameter

Sensor used

Threshold condition

Heart rate

MAX30102

> 120 BPM

SpO₂ level

MAX30102

< 92%

Body motion

MPU6050

High repetitive acceleration

Muscle

activity

EMG sensor

High amplitude muscle

contraction

4. Experimental setup and results

In below Fig. 3 the experimental setup and performance evaluation of our proposed wearable epileptic seizure

detection system. The system’s functionality is tested using real- time motion and physiological data acquisition.

The experimental study consists of four subsections: sensor calibration and data acquisition, hardware

implementation, performance evaluation, and power consumption analysis Database. The MPU6050 motion

sensor was calibrated using a zero-motion reference method along with a six-point orientation test to minimize

drift and improve motion accuracy. Sensor data was sampled at a frequency of 100 Hz, and a low-pass filtering

technique was applied to eliminate noise artifacts and unwanted disturbances from the collected motion

signals. For heart rate monitoring, the pulse sensor was tested under controlled conditions to ensure reliable

and accurate readings. A baseline heart rate was recorded for each subject, and a threshold-based detection

method was implemented to identify abnormal spikes or sudden drops in heart rate that may occur during

seizure episodes. Seizure-like motion patterns were simulated through controlled experimental tests based on

clinically observed tremors and abnormal body movements associated with epileptic seizures.

Fig. 3: Experimental setup for IoT-based human wearable epileptic seizure alert system.

The collected sensor data was continuously logged and processed in real time using the ESP32C3 XIAO

microcontroller. The proposed wearable system was designed as a compact, low-power, and real-time seizure

detection bracelet integrating multiple hardware components for continuous monitoring and emergency

response. The ESP32C3 XIAO microcontroller serves as the central processing unit responsible for sensor data

acquisition, signal processing, and real-time decision-making. The MPU6050 accelerometer and gyroscope

sensor detects abnormal body movements and orientation changes indicative of seizure activity. A pulse sensor

continuously monitors the user’s heart rate and identifies irregular fluctuations that may correlate with seizure

events. In addition, a DF-Player Mini module connected to a speaker generates an audible alert to notify nearby

individuals and provide immediate assistance during emergencies. Fig. 4 represents the circuit connections of

a sensor-based monitoring system using ESP32C3. Various sensors such as MPU6050 and MAX30102 are

interfaced with the microcontroller to collect real-time data. The system also includes a muscle sensor and MP3

module for additional functionalities like signal detection and audio output. All components are interconnected

to enable efficient data processing and monitoring.

Fig. 4: Circuit diagram of seizure alert system.

5. Performance evaluation

The performance of the proposed seizure detection system was evaluated using real-time motion and

physiological data collected under multiple testing conditions. The system was tested in two major scenarios:

seizure simulation and normal daily activities. Seizure simulation experiments were conducted to evaluate the

capability of the system to accurately identify seizure events, while normal activity testing was performed to

analyze system robustness and reduce false positive detections during routine movements.

Experimental results demonstrated that the proposed system achieved a seizure detection accuracy of 92.5%

with a false positive rate of 5.2%, indicating reliable and efficient performance in real-time monitoring

applications. Compared to traditional EEG-based systems and motion-only detection methods, the proposed

multi-sensor wearable system provided improved detection accuracy while maintaining a lightweight,

portable, and user- friendly design. To further enhance performance, two different detection approaches were

analyzed: a generic detection model using fixed threshold values and a personalized detection model using

adaptive thresholds based on individual physiological and movement patterns. The experimental evaluation

revealed that the personalized model significantly outperformed the generic model by improving detection

accuracy by 12.4% and substantially reducing false positive detections. These results highlight the effectiveness

of adaptive thresholding techniques in wearable epileptic seizure detection systems and demonstrate the

importance of personalized monitoring for achieving higher reliability and better patient safety.

From Table 3 it is evident that the proposed system outperforms generic and personalized models in terms of

detection accuracy, false positive reduction, and response time. The integration of multi-sensor data, including

motion, heart rate, SpO₂, and EMG signals, significantly enhances the reliability of seizure detection while

ensuring efficient real-time alert generation.

Table 3: Performance comparison.

Parameter

Generic model

Personalized model

Proposed system

Detection accuracy

80.1%

89.2%

92.5%

False positive rate

14.8%

8.3%

5.2%

Response time

8–10 sec

5–7 sec

3–5 sec

Parameters used

Motion only

Motion + HR

Motion + HR + SpO₂ + EMG

5.1 Power consumption and system lifetime analysis

The power consumption of the system is evaluated by measuring the current draw of each component. The key

observations are:

XIAO ESP32-C3: ~70mA in active mode,

~10mA in deep sleep mode.

MPU6050 (Motion Sensor): ~3.9mA in active mode,

~6µA in sleep mode.

GSM Module (SIM800L/SIM900A): ~100-150mA in idle mode, up to 2A during transmission.

5.2 Push button

Negligible power consumption.

The Node-MCU enters deep sleep mode when no motion is detected.

The GSM module remains in power-saving mode, waking only for SMS transmission. The MPU6050 operates

intermittently to detect critical motion patterns, reducing unnecessary power consumption.

6. Results and discussion

This section presents the experimental evaluation of the proposed IoT-based wearable epileptic seizure alert

system. The system performance is analyzed based on physiological, motion, and muscle activity parameters

collected using the MAX30102, MPU6050, and EMG sensors. Prototype-level testing was conducted under

controlled conditions to validate the effectiveness of the multi-sensor fusion approach.

6.1 Heart rate and SpO₂ analysis

Heart rate and blood oxygen saturation are critical physiological parameters that exhibit noticeable variations

during tonic–clonic seizure events. In the proposed system, these parameters are continuously monitored using

the MAX30102 sensor to identify abnormal physiological behavior. As shown in the Table 4, the heart rate

increases significantly during seizure activity compared to the normal resting state. This sudden rise in heart

rate indicates abnormal cardiovascular response and contributes to seizure detection reliability. Table 5

illustrates a noticeable decrease in SpO₂ levels during seizure events. This reduction reflects respiratory

irregularities that often occur during tonic– clonic seizures, supporting the use of SpO₂ as an important

physiological indicator.

Table 4: Comparison of heart rate values during normal and seizure condition

Condition

Heart Rate (BPM)

Normal

72–85

Seizure

120–150

Table 5: Variation in SpO₂ levels during normal and seizure condition

Condition

SpO₂ (%)

Normal

95–99%

Seizure

88–92%

6.2. Motion and muscle activity analysis

Tonic–clonic epileptic seizures are characterized by sudden, repetitive, and involuntary body movements. To

capture these motion patterns, the proposed system utilizes the MPU6050 sensor, which measures tri-axial

acceleration of the user’s body, as shown in Fig. 5. During normal daily activities, the acceleration values remain

relatively stable and low, whereas seizure events produce high-amplitude and repetitive acceleration spikes.

Tonic–clonic epileptic seizures are characterized by sudden, repetitive, and involuntary body movements. To

capture these motion patterns, the proposed system utilizes the MPU6050 sensor, which measures tri-axial

acceleration of the user’s body. During normal daily activities, the acceleration values remain relatively stable

and low, whereas seizure events produce high-amplitude and repetitive acceleration spikes. Analyzing these

motion patterns enables effective differentiation between normal movement and seizure-induced activity.

Tonic–clonic seizures are characterized by involuntary body movements and muscle stiffness. To capture these

features, the MPU6050 sensor and EMG sensor are employed in the proposed system. Fig. 6 shows an increase

in EMG signal amplitude during seizure activity, representing involuntary muscle contractions and stiffness.

This parameter enhances detection accuracy when combined with motion and physiological data The

experimental results demonstrate that the integration of physiological, motion, and muscle activity parameters

enhances seizure detection reliability. Unlike single parameter systems, the proposed multi-sensor approach

reduces false positives and improves robustness in real- world scenarios. Although the system is evaluated at

the prototype level, the obtained results indicate strong potential for wearable epilepsy monitoring

applications.

Fig. 5: The acceleration pattern recorded by the MPU6050 sensor during epileptic seizure activity

Fig. 6: EMG signal variation indicating muscle stiffness during seizure events.

7. Conclusion

In this paper, we have presented a wearable epileptic seizure detection bracelet designed for real-time

monitoring and emergency alerting. Our experimental evaluation demonstrates that by integrating XIAO

ESP32-C3, MPU6050, and a GPS module, the system effectively detects seizure- related motion patterns and

transmits alerts to caregivers. The system achieves real-time seizure detection while maintaining a battery life

of approximately 14- 16 hours on a 1200mAh battery. By implementing low-power sleep modes, the

operational lifetime can be extended, ensuring efficient long-term usage. Furthermore, the optimized feature

extraction process, using a four-second sliding window with 80% overlap, enables accurate seizure

classification while balancing computational efficiency and power consumption. Additionally, the compact and

lightweight design enhances the wearability of the device, making it a portable and practical solution for seizure

monitoring. The system’s low energy consumption and reduced dependency on Wi-Fi make it a suitable

alternative to hospital- based monitoring, providing an affordable and accessible solution for patients. Overall,

this wearable seizure detection system can significantly improve patient safety by providing timely alerts to

family members, reducing the risks associated with epileptic seizures, and ultimately enhancing the quality of

life for individuals with epilepsy.

Acknowledgement

The authors express their sincere gratitude to the Department of Electronics and Telecommunication

Engineering, Sinhgad Institute of Technology, Lonavala, for providing the necessary facilities and support for

this project. We also thank the faculty members, laboratory staff, and our fellow students for their cooperation

and support. Finally, we acknowledge the support of Savitribai Phule Pune University, Pune, for providing the

academic framework that enabled this research.

CRediT Author Contribution Statement

Dnyaneshwar S. Mantri: Conceptualization, Methodology, Supervision, Writing – Review & editing.

Sushilkumar S. Salve: Project administration, Supervision, Writing – Review & editing. Yogesh Budhalkar:

Data Curation, Methodology, Software, Investigation. Sakshi Bhaltilak: Data curation, Validation, Writing –

Original Draft. Dnyandev More: Validation, Data curation, Software. All authors have read and agreed to the

published version of the manuscript.

Funding Declaration

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-

profit sectors.

Data Availability Statement

No publicly available dataset was used in this study. The work is based on the design, development, and

prototype evaluation of a wearable epilepsy monitoring system. Additional technical details related to the

prototype are available from the corresponding author upon reasonable request.

Conflict of Interest

There is no conflict of interest.

Artificial Intelligence (AI) Use Disclosure

The authors declare that artificial intelligence (AI)-assisted tools were used only for language refinement,

grammar improvement, and manuscript structuring purposes during the preparation of this work. All technical

content, experimental implementation, results, and interpretations were independently developed and verified

by the authors.

Supporting Information

Not applicable.

References

[1]

D. Sopic, A. Aminifar, D. Atienza, e-Glass: A wearable system for real-time detection of epileptic seizures,

in Proceedings of the 2018 IEEE International Symposium on Circuits and Systems (ISCAS), IEEE, 2018,

1–5, doi: 10.1109/ISCAS.2018.8351728.

[2]

S. Jahan, F. Nowsheen, M. M. Antik, M. S. Rahman, M. S. Kaiser, A. S. M. S. Hosen, I. H. Ra, AI-based epileptic

seizure detection and prediction in internet of healthcare things: A systematic review, IEEE Access,

2023, 11, 30690–30725, doi: 10.1109/ACCESS.2023.3251105.

[3]

W. Truccolo, O. J. Ahmed, M. T. Harrison, E. N. Eskandar, G. R. Cosgrove, J. R. Madsen, A. S. Blum, N. S.

Potter, L. R. Hochberg, S. S. Cash, Neuronal ensemble synchrony during human focal seizures, Journal of

Neuroscience, 2014, 34, 9927-9944, doi: 10.1523/JNEUROSCI.4567-13.2014.

[4]

J. Tang, R. E. Atrache, S. Yu, U. Asif, M. Jackson, S. Roy, M. Mirmomeni, S. Cantley, T. Sheehan, S. Schubach,

C. Ufongene, S. Vieluf, C. Meisel, S. Harrer, T. Loddenkemper, Seizure detection using wearable sensors

and machine learning: Setting a benchmark, Epilepsia, 2021, 62, 1807-1819, doi: 10.1111/epi.16967.

[5]

A. Pantelopoulos, N. G. Bourbakis, A survey on wearable sensor-based systems for health monitoring

and prognosis, IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews),

2010, 40, 1-12, doi: 10.1109/TSMCC.2009.2032660.

[6]

J. Dan, B. Vandendriessche, W. V. Paesschen, D. Weckhuysen, A. Bertrand, Computationally-efficient

algorithm for real-time absence seizure detection in wearable electroencephalography, International

Journal of Neural Systems, 2020, 30, doi: 10.1142/S0129065720500355.

[7]

J. Zhang, L. Swinnen, C. Chatzichristos, V. Broux, R. Proost, K. Jansen, B. Mahler, N. Zabler, N. Epitashvilli,

M. Dümpelmann, A. Schulze-Bonhage, E. Schriewer, U. Ermis, S. Wolking, F. Linke, Y. Weber, M.

Symmonds, A. Sen, A. Biondi, M. P. Richardson, A. Sulaiman I, A. I. Silva, F. Sales, G. Vértes, W. V.

Paesschen, M. D. Vos, Multimodal Wearable EEG, EMG and accelerometry measurements improve the

accuracy of tonic-clonic seizure detection in-hospital, arXiv preprint, 2024, doi:

10.48550/arXiv.2403.13066.

[8]

J. E. Chukwunazo, O. O. Remigius, S. K. Ogbuokebe, Design and simulation of microcontroller based

wireless patient intelligent health monitoring system with GSM alert technology, International Journal

of Engineering Trends and Technology (IJETT), 2015, 24, 87-95, doi: 10.14445/22315381/IJETT-

V24P217.

[9]

W. Zhao, W. F Wang, L. M. Patnaik, B. C. Zhang, S. J. Weng, S. X. Xiao, D. Z. Wei, H. F. Zhou, Residual and

bidirectional LSTM for epileptic seizure detection, Frontiers in Computational Neuroscience, 2024, 18,

doi: 10.3389/fncom.2024.1415967.

[10]

P. Grant, M.Z. Islam, Seizure Detection from Electroencephalogram Signals via Wavelets and Graph

Theory Metrics, arXiv preprint, 2023, doi: 10.48550/arXiv.2312.00811.

[11]

A. Shoeb, J. Guttag, Application of machine learning to epileptic seizure detection, In Proceedings of the

27th International Conference on International Conference on Machine Learning (ICML'10),

Omnipress, 2010, 975–982, doi: 10.5555/3104322.3104446.

[12]

T. Conlon, H.J. Ruskin, M. Crane, Seizure characterisation using frequency-dependent multivariate

dynamics, Computers in Biology and Medicine, 2009, 39, 760-767, doi:

10.1016/j.compbiomed.2009.06.003.

[13]

M. Z. Poh, T. Loddenkemper, C. Reinsberger, N.C. Swenson, S. Goyal, J.R. Madsen, R.W. Picard, Autonomic

changes with seizures correlate with postictal EEG suppression, Neurology, 2012, 78, 1868-1876, doi:

10.1212/WNL.0b013e318258f7f.

[14]

E. Pereira, An IoT Approach to Personalised Remote Monitoring and Management of Epilepsy, 2017

14th International Symposium on Pervasive Systems, Algorithms and Networks & 2017 11th

International Conference on Frontier of Computer Science and Technology & 2017 Third International

Symposium of Creative Computing, 2017, doi: 10.1109/ISPAN-FCST-ISCC.2017.34.

[15]

E. Donner, O. Devinsky, D. Friedman, Wearable digital health technology for epilepsy, New England

Journal of Medicine, 2024, 390, 736-745, doi: 10.1056/NEJMra2301913.

[16]

A. Bernini, J. Dan, P. Ryvlin, Ambulatory seizure detection, Current opinion in neurology, 2024, 37, 99-

104, doi: 10.1097/WCO.0000000000001248.

[17]

M. Sasseville, E. Attisso, M. P. Gagnon, J. M. W. Supper, S. Ouellet, S. Amil, E. B. Assi, D. Khoa Nguyen,

Performance, impact and experiences of using wearable devices for seizure detection in community-

based settings: a mixed methods systematic review, Mhealth, 2024, 10, 27, doi: 10.21037/mhealth-24-

[18]

A. Ahuja, S. Agrawal, S. Acharya, N. Batra, V. Daiya, N. Batra, Advancements in wearable digital health

technology: a review of epilepsy management, Cureus, 2024, 16, doi: 10.7759/cureus.57037.

[19]

J. Jeppesen, J. Christensen, O. A. Petersen, S. Fenger, S. A. Larsen, S. Wüstenhagen, S. R. Wagner, P.

Johansen, S. Beniczky, Seizure detection using wearable electrocardiogram connected to a smartphone:

a phase 3 clinical validation study, eBioMedicine, 2025, 116, 1228, doi: 10.1038/s41597-025-05580-x.

[20]

MPU-6050 Product Specification, InvenSense, 2013, https://www.digikey.in/en/products/detail/tdk-

invensense/MPU-6050/4038009, Accessed 20 April 2026.

[21]

G. Regalia, F. Onorati, M. Lai, C. Caborni, R. W. Picard, Multimodal wrist-worn devices for seizure

detection and advancing research: Focus on the Empatica wristbands, Epilepsy Research, 2019,

153, 79-82, doi: 10.1016/j.eplepsyres.2019.02.007.

[22]

L. Faust, J. Cui, C. Knepper, M. Nasseri, G. Worrell, B. H. Brinkmann, Detecting diverse seizure types with

wrist-worn wearable devices: a comparison of machine learning approaches, MDPI Sensors, 2025, 25,

5562, doi: 10.3390/s25175562.

[23]

G. Japaridze, D. Loeckx, T. Buckinx, S. Armand Larsen, R. Proost, K. Jansen, P. MacMullin, N. Paiva, S.

Kasradze, A. Rotenberg, L. Lagae, Automated detection of absence seizures using a wearable

electroencephalographic device: a phase 3 validation study and feasibility of automated behavioral

testing, Epilepsia, 2023, 64, S40-S46, doi: 10.1111/epi.17200.

[24]

J. Tang, R. El Atrache, S. Yu, U. Asif, M. Jackson, S. Roy, M. Mirmomeni, S. Cantley, T. Sheehan, S. Schubach,

C. Ufongene, S. Vieluf, C. Meisel, S. Harrer, T. Loddenkemper, Seizure detection using wearable sensors

and machine learning: Setting a benchmark, Epilepsia, 2021, 62, 1807–1819, doi: 10.1111/epi.16967.

[25]

MAX30102 Heart Rate and Pulse Oximeter Sensor Datasheet, Maxim Integrated, 2020,

https://www.scribd.com/document/902561072/MAX30102-DATASHEET, Accessed 20 April 2026.

[26]

C. Baumgartner, J. Baumgartner, C. Lang, T. Lisy, J. P. Koren, Seizure Detection Devices, Journal of

Clinical Medicine, 2025, 14, 863, doi: 10.3390/jcm14030863.

[27]

A. Y. Adwitiya, D. H. Hareva, I. A. Lazarusli, Epileptic Alert System on Smartphone, 2017 International

Conference on Soft Computing, Intelligent System and Information Technology (ICSIIT), 2017, 288-291,

doi: 10.1109/ICSIIT.2017.59.

[28]

Espressif Systems, ESP32-C3 Series Datasheet, 2022

https://www.espressif.com/sites/default/files/documentation/esp32-c3_datasheet_en.pdf, Accessed

21 April 2026.

[29]

F. Samie, L. Bauer, J. Henkel, IoT technologies for embedded computing: a survey, In Proceedings of the

Eleventh IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System

Synthesis (CODES '16), Association for Computing Machinery, 2016, 1–10, doi:

10.1145/2968456.2974004.

[30]

A. Agrahri, A. Tyagi, D. Kumar, S. Kusmakar, M. Palaniswami, B. Yan, Detection of Epileptic Seizure Using

Accelerometer Time Series Data and Hidden Markov Model, 2022 44th Annual International

Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2022, 2426-2429, doi:

10.1109/EMBC48229.2022.9871914.

[31]

SIMCom, SIM800L GSM Module Hardware Design, SIMCom Wireless Solutions, 2020,

https://components101.com/wireless/sim800l-gsm-module-pinout-datasheet-equivalent-circuit-

specs, Accessed 20 April 2026.

[32]

S. Beniczky, I. Conradsen, O. Henning, M. Fabricius, P. Wolf, Automated real-time detection of tonic-

clonic seizures using a wearable EMG device, Neurology, 2018, 90, 428–434, doi:

10.1212/WNL.0000000000004893.

[33]

u-blox AG, NEO-6 u-blox 6 GPS Modules Data Sheet, 2011, https://content.u-

blox.com/sites/default/files/products/documents/NEO-6_DataSheet_(GPS.G6-HW-09005).pdf,

Accessed 27 April 2026.

[34]

I. Singh, D. Kumar, Improving IOT Based Architecture of Healthcare System, 2019 4th International

Conference on Information Systems and Computer Networks (ISCON), 2019, 113-117, doi:

10.1109/ISCON47742.2019.9036287.

[35]

T. M. Ingolfsson, U. Chakraborty, X. Wang, S. Beniczky, P. Ducouret, S. Benatti, P. Ryvlin, A. Cossettini, L.

Benini, EpiDeNet: An Energy-Efficient Approach to Seizure Detection for Embedded Systems, arXiv

preprin, 2023, doi: 10.48550/arXiv.2309.07135.

[36]

S. Mekruksavanich, W. Phaphan, A. Jitpattanakul, Epileptic seizure detection in EEG signals via an

enhanced hybrid CNN with an integrated attention mechanism, Mathematical Biosciences and

Engineering, 2025, 22, 73-105, doi: 10.3934/mbe.2025004.

[37]

I. Y. Potter, G. Zerveas, C. Eickhoff, D. Duncan, Unsupervised Multivariate Time-Series Transformers for

Seizure Identification on EEG, arXiv preprint, 2023, doi: 10.1109/ICMLA55696.2022.00208.

[38]

Z. ElSayed, M. Ozer, N. Elsayed, A. Abdelgawad, Machine Learning Based IoT Adaptive Architecture for

Epilepsy Seizure Detection: Anatomy and Analysis, arXiv preprint, 2023, doi:

10.48550/arXiv.2305.19347.

[39]

S. Aziz, A. A. M. Ali, H. Aslam, N. U. Ain, A. Tariq, Z. Sohail, S. Murtaza, Wearable Artificial Intelligence

for Epilepsy: Scoping Review, Journal of medical Internet research, 2025, 27, e73593, doi:

10.2196/73593.

[40]

L. Hadady, P. Klivényi, D. Fabó, S. Beniczky, Wearable Devices for Seizure Detection: Practical

Experiences and Recommendations, Epilepsia, 2018, 3, 2–16, doi: 10.1111/epi.17189.

[41]

A. Djemal, D. Bouchaala, A. Fakhfakh, O. Kanoun, Wearable Electromyography Classification of Epileptic

Seizures: A Feasibility Study, Bioengineering, 2023, 10, 703, doi: 10.3390/bioengineering10060703.

[42]

W. T. Kerr, K. N. McFarlane, G. F. Pucci, The present and future of seizure detection, prediction, and

forecasting with machine learning, including the future impact on clinical trials, Frontiers in neurology,

2024, 15, doi: 10.3389/fneur.2024.1425490.

[43]

M. Bhagubai, C. Chatzichristos, L. Swinnen, J. Macea, J. Zhang, L. Lagae, K. Jansen, A. Schulze-Bonhage,

F. Sales, B. Mahler, Y. Weber, W. V. Paesschen, M. De Vos, SeizeIT2: Wearable dataset of patients with

focal epilepsy, Scientific Data, 2025, 12, 1228, doi: 10.1038/s41597-025-05580-x.

Publisher Note: The views, statements, and data in all publications solely belong to the authors and

contributors. GR Scholastic is not responsible for any injury resulting from the ideas, methods, or products

mentioned. GR Scholastic remains neutral regarding jurisdictional claims in published maps and institutional

affiliations.

Open Access

This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which

permits the non-commercial use, sharing, adaptation, distribution and reproduction in any medium or format,

as long as appropriate credit to the original author(s) and the source is given by providing a link to the Creative

Commons License and changes need to be indicated if there are any. The images or other third-party material

in this article are included in the article's Creative Commons License, unless indicated otherwise in a credit line

to the material. If material is not included in the article's Creative Commons License and your intended use is

not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly

from the copyright holder. To view a copy of this License, visit: https://creativecommons.org/licenses/by-

nc/4.0/