Performance analysis of cognitive radio networks over к-μ fading channel with noise uncertainty

Abstract

In order to be able to perform a Spectrum Sensing in a reliably manner we must take in consideration a physical fading model which will fit with the reality presented. In this work we analyze the energy detection characteristics over a generalized fading channel, modeled by the κ-μ distribution. Admitting the noise power estimation error, the worst-case of the probabilities of miss detection and false-alarm, due to the noise uncertainty, are derived under the spectrum utilization constraint. An optimal threshold selection is presented in order to allow the energy detector to operate in regions of low SNR. Field measurements are used to investigate in practice the usefulness of the κ-μ fading channel in cooperative spectrum sensing scenarios. Comparisons are performed against Rice fading model where it is possible to notice a great advantage in using the κ-μ distribution to describe the fading channel.