Challenges and opportunities to location independent human activity recognition utilizing Wi-Fi sensing

Wireless sensing has emerged as a dynamic field with diverse applications across smart cities, healthcare, the internet of things (IoT), and virtual reality gaming. This burgeoning area capitalizes on the capacity to detect locations, activities, gestures, and vital signs by assessing their impact on ambient wireless signals. This review critically examines the prevailing challenges within wireless sensing and predicts future research trajectories. Even with the potential for nuanced signal processing facilitated by Wi-Fi propagation, its efficacy is impeded by noise interference in confined areas during transmission and reception. Consequently, this work aims to augment signal processing performance accuracy by delving into the most promising techniques and underexplored methods utilizing channel state information (CSI). Furthermore, the work offers a view into the potential of human activity recognition predicated on CSI properties. The study focusses on exploring location-independent sensing technique based on CSI, discussing relevant considerations and contemporary approaches used in Wi-Fi sensing tasks. The optimal practices in analysis are based on model design, data collection, and result interpretation. The discussions analysis investigates in detail the representative applications and outlines the major considerations of developing human activity recognition human activity recognition (HAR) based on Wi-Fi by analyzing the current critical issues of CSI-based behavior recognition methods and pointing out possible future research directions.