With the increasing need for wireless data transfer, Wi-Fi networks have rapidly grown in recent years providing high throughput and easy deployment. At present, Access Points (APs) can be found easily wherever we go, thus Wi-Fi sensing applications have caught a great deal of interest from the research community. Since human presence and movement influence the Wi-Fi signals transmitted by APs, it is possible to exploit those signals for human detection, localization, pose estimation and, even, re-identification (Re-ID) tasks. Traditional techniques for Wi-Fi sensing applications are usually based on the Received Signal Strength Indicator (RSSI) measurement. However, recently, due to the RSSI instability, the researchers in this field propose more and more Channel State Information (CSI) measurement based methods. Indeed, combining Wi-Fi signal CSI measurements with Machine Learning algorithms, according to the state-of-the-art, is a promising strategy. The proposed work aims to deploy an automatic Device-Free Wi-Fi system for human detection, localization, pose estimation and, especially, Re-ID tasks based on CSI processing and Deep Learning strategies. Due to the lack of Device-Free Wi-Fi based human Re-ID related works in the current state-of-the-art, the acquisition of a dataset fitting all the afore mentioned tasks is planned.
Nowadays, the scientific community has highlighted the Channel State Information (CSI) measurement validity for Wi-Fi sensing systems in the fields of human detection, localization and pose estimation. Such measurement, however, may still be used to extract new features not considered yet and may be integrated to different data types or signals to develop new Wi-Fi sensing applications or enhance existing ones. For example, combining the CSI, which is the physical layer information, with upper layer information such as Medium Access Control (MAC), Transmission Control Protocol (TCP) and Internet Protocol (IP) could help enhance the Wi-Fi sensing applications in terms of robustness and generalization. Again, the CSI may be combined with audio or video signals to provide high performance and less human efforts of training Machine Learning algorithms [1]. In addition, the state-of-the-art lacks Device-Free Wi-Fi re-identification (Re-ID) methods and, consequently, of a dataset fitting this purpose. Performing research activity in the field of Device-Free Wi-Fi Re-ID task then not only provides a new research branch for Wi-Fi sensing but, consequently, would provide a dataset suited for the task supporting and encouraging the growth of the state-of-the-art in this field.
[1] M. Zhao et al., "Through-Wall Human Pose Estimation Using Radio Signals," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7356-7365, 2018.