Desempenho do método CF-CPSC em redes de rádios cognitivos sob erros no canal de controle

Abstract

The spectrum sensing technique for cognitive radio networks, called CPSC (Cooperative Power Spectrum Density Split Cancellation) was recently suggested and it has been a model for continuous studies and improvements. It has some advantages like being a robust technique against dynamic noise, and also low computational complexity. From this technique, the CF-CPSC (Circular Folding CPSC) was developed, whose performance expressively overcomes the original. In this dissertation, an analysis about CF-CPSC performance was performed, over report channel errors, making a comparison between two methods. First, sendind only samples and after sending test statistics, by cognitive radios to the center fusion. The performance was also compared and evaluated applying error correcting codes, in the way of extracting best results, always observing their impact at the report channel occupation. Another contribution was developing a novel decision fusion periodogram-based algorithm for centralized cooperative spectrum sensing. Results showed the new method is sensitive to errors, however even with error correcting codes applied, the report channel traffic is smaller, leading to the conclusion that a case-by-case analysis is necessary to define wich is the best model to every situation. Altough, the decision fusion method might be applied to cases that require low data traffic at the report channel, keeping high performance.