ROBUST SPECTRUM SENSING FOR COGNITIVE RADIO

This paper studies spectrum sensing in the context of cognitive radio. The proposed detector is robust with respect to disturbing impulses, that are in practice present in many cases in addition to the Gaussian noise. The presence of impulsive noise deteriorates the performance of the detectors derived using Gaussian assumption. In the paper we model the noise explicitly consisting of two components and derive the proper detector. Asymptotic analysis of the detector is then presented and formulae for probabilities of correct detection and fault alarm are derived. The theoretical findings are verified in our simulation study.