RFALL: Multidirectional Fall Detection and Pseudo-Localization Using RFID Tag Arrays

This study presents RFALL, a contactless monitoring system using ultrahigh-frequency (UHF) radio frequency identification (RFID) tag arrays for directional fall detection with pseudo-localization. The system processes raw received signal strength (RSSI) data with minimal preprocessing (normalization only) to classify static activities (no-activity, standing, and leaning) and dynamic falls (left, right, forward, and backward) while providing column-based spatial positioning within monitoring zones. The system was validated at distances of 3.5–5.5 m across three different wall configurations, achieving a maximum precision of 97.6% using machine learning (ML) algorithms, with a support vector machine showing superior performance. The system’s ability to utilize unique RFID tag IDs enables simultaneous monitoring of multiple individuals, enhancing its applicability in patient care and management while maintaining cost-effectiveness through minimal hardware requirements.