Abstract
Spectrum sensing in cognitive radio (CR) is an emerging technologywhich enables coexistence of systems; thus, the precious resourcesof frequency bands can be utilized more efficiently. In this paper,narrowband interference (NBI) detection in an orthogonal frequencydivision multiplexing (OFDM) based cognitive radio system isaddressed. According to how much available information about NBI isknown to the detector, several detection algorithms based on theNeyman-Pearson philosophy are proposed. In specific, the amount ofreceiver's knowledge about NBI is classified into categoriesaccording to whether or not a) the received power, b) the occupiedfrequency bands, and c) the second moment descriptions of NBI areavailable. In dealing with unknown parameters in the hypothesistesting, generalized likelihood ratio test (GLRT) is employed.Performance analysis is carried out for all proposed algorithms interms of the receiver operating characteristic (ROC). Extensivecomputer simulations are run to verify the accuracy of the analysis.