Spectrum sensing methods and their 6 performance

Abstract

Spectrum sensing is the process of determining if a spectrum slot is occupied or not by a primary signal. This tutorial emphasizes energy detection based spectrum sensing and provides a broad overview of the tools necessary for performance analysis of several spectrum sensing algorithms. A detailed description of the spectrum sensing problem is provided as a binary hypothesis test. The main parameters of interest - decision statistic, detection, and false-alarm probabilities and the decision threshold - are discussed. These parameters of the energy detector, which computes the energy of the received signal, are described. The use of the central limit theorem (CLT) to achieve energy detection with prescribed performance level is discussed. The receiver operating characteristic (ROC) curve and area under the curve (AUC) are described. Fading, a fundamental wireless channel impairment, can be mitigated with multiple antenna techniques, which provide spatial diversity gains. The performance of the energy detector with two low-complexity diversity techniques is described. The performance is analyzed for Rayleigh fading, for spatial correlation, and in the high signal-to-noise ratio (SNR) regime. General analytical techniques are highlighted. Double-threshold energy detector, P-norm detector, and energy detection for full-duplex nodes are described. Alternative to energy detection includes cyclostationary detection, matched filter-based detection, and waveform-based detection. These methods are briefly discussed. Spectrum sensing is an essential part of smart grid, Internet of things, and cognitive radio. An overview is provided.