Separação cega de fontes aplicada no sensoriamento do espectro em rádio cognitivo

Abstract

Cognitive radio technology has been an important area of research in telecommunications for solving the problem of spectrum scarcity. That\'s because in addition to allowing dynamic allocation of the electromagnetic spectrum, cognitive radios must be able to identify the non cognitive user\'s transmission on the channel. This operation is only possible through the continuous sensing of the electromagnetic spectrum. In this context, this paper presents a detailed study on spectrum sensing, an important stage in cognitive radio technology. For the presentation of this work, a detailed study on software dened radio (SDR) was carried out, without which it would be impossible to work with cognitive radios, once they are implemented by means of SDR technology. It was also presented the tools GNU Radio and USRP, which together form a solution of SDR, through implementation of AM receivers. The theoretical foundations of spectrum sensing and blind source separation (BSS) are presented and then is made a detailed study of the use of BSS for spectral sensing. From the study of BSS, it is possible to use new metrics for decision making about the presence or the absence of a primary user in the channel. Throughout the study, simulations and implementations were conducted on MATLAB in order to perform various situations, and, nally, it is presented outcomes and conclusions reached during the work.