A Survey on Human Profile Information Inference via Wireless Signals

Due to the ubiquitous deployment of wireless infrastructures, the radio signal nature of invisibility, and the elimination of the line-of-sight requirement, it has drawn increasing attention in both academia and industry to infer various human-motion related sensitive information, called human profile information (HPI). The basic idea of these techniques is that varying human profiles (i.e., physiological characteristics and motion patterns) may lead to unique and subtle disturbances in environmental wireless signals, which can be then measured and processed to learn HPI. In this survey, we comprehensively review different categories of existing studies based on (1) how they quantize the motion-induced wireless disturbances in concrete wireless measurements; (2) the signal processing techniques for building wireless HPI inference systems; and (3) practical applications that take advantage of inferred HPI. Also, the survey discusses the emerging challenges and future directions on wireless HPI inference.