Wi-BioScan: Human Identification Based on Radio Shadows

摘要

Subject identification is essential in smart environments. Identification of differently disguised human subjects through visual surveillance methods is extremely challenging. Subject may use various artifacts to defy methods based on body shape and size matching. The radio imaging through wireless sensing is ubiquitous, device-free and privacy-preserving. This paper is based on our observation when a subject is placed in-between the line-of-sight of a closely placed pair of wireless nodes. A radio shadow is observed at the receiver due to interaction of the subject with signals. We observed that the subject can be identified by comprehensive analysis of the received radio shadow and its derivative signatures. Wi-Bioscan is the first ever system to find uniqueness among human radio shadows to identify the subject even if it is disguised in four different ways. During our investigation, we evaluated Wi-BioScan system for different indoor/outdoor locations. System is ubiquitous and scalable due to its feature of location independence and generation of a rich template. The method is also robust against surrounding dynamics such as other human presence and object movements.