Periodic Spectrum Sensing Performance and Requirements for Detecting Legacy Users with Temporal and Noise Statistics in Cognitive Radios

In this paper we consider the spectrum sensing performance and requirements for detecting legacy users (LU) in cognitive radios (CR) with periodic scanning. The performance and requirements are studied based on the temporal spectral occupancy statistics of the LU and the sensing signal to noise ratio levels in order to achieve a certain level of detection probability. We model the temporal statistics of the LU as a Poisson Pareto burst process (PPBP) describing a typical WEB service application and the noise as an additive white Gaussian noise (AWGN) process. Theoretical expressions are computed for the detection probabilities, for energy based detection, and we derive expressions for the sensing time requirements to successfully detect the LU based on the temporal and the noise statistics for a given detection criteria.