Quality-Of-Service Prediction For Physical-Layer Security Via Secrecy Maps

While most of the theoretical aspects of physical layer security are well understood, practical applications lag substantially behind theoretical advances. As a step towards the integration of physical-layer security aspects in the radio access system design, the concept of secrecy maps has been recently introduced. Building upon this concept, in this work we focus on system design-related aspects which consider physical-layer security as a service, together with thereby associated secrecy-related Quality-of-Service (QoS) requirements. We adopt a statistical learning framework to characterize the wireless environment from the perspective of semantic security via QoS maps that indicate locations in the wireless environment that surmount a QoS-related performance threshold with a guaranteed level of confidence. To provide secrecy characterization of a particular wireless environment, we propose to learn any-to-any QoS maps by leveraging the tensor completion framework. We demonstrate the advantage of the proposed method over a baseline algorithm based on tensor rank minimization via numerical simulations.