A Multi-Sensor Wearable System for Real-Time Epileptic Seizure Detection Using Machine Learning

Abstract

Epileptic seizures occur suddenly and can lead to severe health risks if not detected early. Continuous and real-time monitoring is essential for timely intervention; however, most existing systems rely on hospital-based EEG setups that are costly and impractical for daily use. This paper presents an AI-based wearable device for early seizure detection using multiple physiological sensors, including EEG, heart rate, galvanic skin response (GSR), and body movement. Sensor data is processed in real time to identify seizure-related patterns. Upon detecting abnormal activity, it triggers immediate alerts via wireless communication to notify caregivers. The wearable solution is portable, energy-efficient, and suitable for continuous monitoring, achieving an accuracy of approximately 88–92%, improving detection reliability while reducing false alarms and enhancing patient safety.