Communication over Random Fields: A Statistical Framework for Cognitive Radio Networks

The performance of cognitive radio is determined by the spectrum occupancy, which is correlated in space, time and frequency. In this paper, the theory of random fields is applied to model the spatially correlated spectrum occupancy. The entire cognitive radio networks are modeled as a plenary grid or a straight line. Two dimensional and one dimensional Ising models are applied to model the joint distributions of spectrum occupancies, respectively. The phase transition in the network connectivity is studied using the random field model. The average delay and jitter are analyzed using the Ising models. Numerical simulations show that the Ising model can effectively predict the connectivity, average delay and jitter.