Unveiling the Hidden Assumptions of Energy Detector Based Spectrum Sensing for Cognitive Radios

Abstract

Cognitive radio is a promising solution to current problem of spectrum scarcity. It relies on efficient spectrum sensing. Energy detection is the most dominantly used spectrum sensing approach owing to its low computational complexity and ability to identify spectrum holes without requiring a priori knowledge of primary transmission characteristics. This paper offers a comprehensive tutorial on energy detection based spectrum sensing and presents an in depth analysis of the test statistic for energy detector. General structure of the test statistic and corresponding threshold are presented to address existing ambiguities in the literature. The derivation of exact distribution of the test statistic, reported in the literature, is revisited and hidden assumptions on the primary user signal model are unveiled. In addition, the scope of detection probability results is discussed for identifying various classes of random primary transmissions. Gaussian approximations of the test statistic are investigated. Specifically, the roles of signal to noise ratio and performance constraint in terms of probability of detection or false alarm are highlighted when Normal approximations are used in place of exact expressions.