WiFiWave: Human Activity Recognition through Wi-Fi Sensing: WiFiWave

As the global population continues to expand and age 60 and beyond. People's ability to perform routine tasks and physical activity levels decrease with age, which influences their physical and psychological well-being. Proposed research addresses the evolving challenges of an aging global population, emphasizing the need for advanced in-home monitoring systems. Focused on Human Activity Recognition (HAR), the study explores the modules of HAR systems and classifies data sources into environmental and wearable sensors. It advocates for Radio Frequency (RF)-based solutions, particularly Wi-Fi CSI, due to their non-intrusive nature, privacy preservation, and effectiveness. The proposed research introduces the UT-HAR dataset for human activity recognition through WiFi sensing and classification module utilizes Residual Neural Networks (ResNet) and Convolutional Neural Networks (CNN) for activity classification. Key performance indicators are used to assess the suggested models, with a focus on AUC-ROC, accuracy, precision, and recall. Findings underscore the potential of Wi-Fi CSI-based systems for accurate and non-intrusive monitoring of elderly individuals, contributing to advancements in healthcare technology.