Distributed estimation for Radio Environment Map in cognitive radio networks

Radio Environment Map (REM) is a map that indicates the radio signal strength (RSS) over a finite geographical region. It enables cognitive radio (CR) users to opportunistically access to the spectrum unoccupied by the primary users (PUs) who are licensed to use specific spectrum. In this paper, a basis expansion based scheme is presented to estimate REM. We take the sparsity of preselected basis functions and the inherent time-varying spectrum usage of primary users into account such that the RSS estimation is formulated as a recursive least squares (RLS) problem with an l1 norm constraint. A distributed algorithm is presented to solve the RLS problem. Each SU only need to exchange information with its one-hop neighbours to sketch the REM with radio signal strength. Finally, simulation studies are presented to validate the performance of the proposed method.