A team of seven students from Helwan University's International Technological College has created a groundbreaking project called WiSafe. This system utilizes WiFi signals to detect falls among seniors, providing a non-intrusive and cost-effective solution for caregivers and families. The project, supervised by Dr. Simon Ezzat, Dr. Mohamed Sabah, and Engineer Nadira Helal, aims to offer a reliable and efficient way to monitor seniors' safety.
The WiSafe system works by analyzing changes in WiFi signals caused by human movement within a room. Using WiFi Channel State Information (CSI) technology, the system can detect falls and alert caregivers or family members instantly. This approach eliminates the need for cameras or wearable devices, addressing concerns about privacy and user comfort. The team collected and processed data to develop and train AI models, achieving a 96.4% recall accuracy for fall detection.
The WiSafe system consists of a presence detection model and a unified human activity recognition and fall classification model. The presence detection model determines whether a person is in the room, while the unified model recognizes and classifies human activities, including falls, sitting, standing, and walking. The system uses a FastAPI-based backend and a Flutter-developed mobile application to provide real-time alerts to caregivers.
The project's development involved several stages, including data collection, signal processing, and AI model training. The team used data from CSI-Bench, a comprehensive dataset containing over 334,276 files and 59.8 GB of data. By integrating academic knowledge with practical training, the team created a comprehensive system that addresses a significant challenge in elderly care.
According to Dr. Osama Al-Qabisi, academic supervisor of the project, WiSafe demonstrates the potential of Helwan University's students to develop innovative solutions that integrate academic knowledge with practical applications. Dr. Sayed Qandeel, president of Helwan University, emphasized that such projects reflect the university's commitment to providing education that enables students to transform knowledge into real-world solutions.
The WiSafe system has shown promising results, achieving a 96.25% F1 score for fall detection and 93.2% accuracy for general activity classification. The system's response time is less than 3 seconds, ensuring timely alerts to caregivers. With its non-intrusive and cost-effective approach, WiSafe has the potential to significantly improve elderly care and provide peace of mind for families and caregivers.
Key points
- The WiSafe system uses WiFi signals to detect falls among seniors, eliminating the need for cameras or wearable devices.
- The system achieved a 96.4% recall accuracy for fall detection and 93.2% accuracy for general activity classification.
- The project's development integrated academic knowledge with practical training, demonstrating the potential of Helwan University's students to develop innovative solutions.
