Wi-Fi CSI-based Human Presence Detection Using DTW Features and Machine Learning
Julio C. H. Soto, Iandra Galdino, Brenda G. Gouveia, Egberto Caballero, Vinícius Ferreira, Débora C. Muchaluat-Saade, Célio Albuquerque
Abstract
With the development of smart devices, human detection and localization became important tasks for several applications including security, healthcare monitoring, entertainment, and so on. Existing signal-based detection systems, mostly focus on detecting human activities and classifying them by Machine Learning (ML) methods, like Support Vector Machine (SVM) and Random Forest (RF). This paper focuses on device-free presence detection. We propose a specific setup for collecting Wi-Fi based Channel State Information (CSI) data for detecting human presence. The proposal includes the application of Dynamic Time Warping (DTW) algorithm features to compare the differences between empty rooms and filled rooms. The proposed architecture and approach achieves competitive accuracy when compared to the existing technologies.