Deep sensing for 5G spectrum sharing: A random finite set approach

In this paper, a new detection framework, namely, deep sensing (DS), is proposed for 5G spectrum sharing, which is designed to proactively recover some informative states associated with realistic cognitive links (e.g., fading gains), except for detecting the occupancy of primary-band. Relying on a dynamic state-space approach, a unified mathematical model is formulated. The Bernoulli random finite set (BRFS) is exploited to theoretically characterize the complex DS procedures. A Bernoulli filter algorithm is suggested to recursively estimate unknown PU states accompanying related link information, which is further implemented by particle filtering. The proposed DS algorithm is applied to detect primary users over more challenging time-varying fading channels. Numerical simulations validate the new scheme. Spectrum sensing can be effectively implemented by estimating time-varying fading gains jointly.