Wi-Fi CSI-Based Fall Detection Using CNN with Advanced Decision Mechanism

This paper introduces a novel approach, denoted ADM-Enhanced CNN, to Fall Detection (FD) using Wi-Fi Channel State Information (CSI) through a specialized Convolutional Neural Network (CNN) enriched with an Advanced Decision Mechanism (ADM). Our methodology, validated using a “left-out user” strategy, emphasizes the model's generalization capability. A comparative analysis with literature demonstrates the superiority of our ADM-Enhanced CNN, highlighting its performance in real-time, accurate fall detection.