Sensoriamento espectral cooperativo com fusão de autovalores para sinais OFDMA e outros sinais de banda larga

Abstract

A possible solution for the spectral scarcity problem is the opportunistic frequency spectrum utilization. The art of verifying the spectrum in pursuit of available positions for transmission occupies a prominent space for cognitive radio development. Aiming at increasing the efficiency of the spectrum sensing process by diminishing the delay time, wideband spectrum sensing techniques are being studied and enhanced. This dissertation presents a new spectrum sensing technique based on the eigenvalues of the covariance matrix of the received signal by cognitive radios. It is also proposed a new method of data fusion through the combination of the eigenvalues of the covariance matrix computed by each cooperating cognitive radio. The test statistics for the eigenvalue-based spectrum sensing of narrowband signals are empirically modified. Simulation results have shown that the proposed combination method outperforms other combination methods such as sample fusion and decision fusion for any test statistic. An adaptation of the spectrum sensing technique to detect the occupancy state of subchannels in OFDMA signals is also proposed. Ultimately, an analysis of the data volume over the control channel shows a reduction of the amount of information transmitted to the fusion center with the proposed technique relative to the sample fusion.