Redes cognitivas: Caracterização do espectro usando redes bayesianas

Resumen

This article presents the use of Bayesian networks and probabilistic inference in cognitive radios, applied to the characterization of the electromagnetic spectrum, allowing the adaptation of this characterization in function of specific local and temporal characteristics. Unifying data states (zero or one) and the states of the spectrum (presence of primary user, noise, silence etc), Bayesian inference associated with the proposed model of cognition, provides a range of information about the spectrum that serves both for the MAC layer as for the control layer, serving to adjust the cognitive radio parameters according to statistics from the spectrum behavior.