Candidate spectral estimation for cognitive radio

Abstract

Introducing a different view point of traditional filter bank approach for spectral estimation it is derived a spectral estimation method able to detect a given spectral shape forming part or contributing to a given data record. The method provides an estimated power level of the spectral shape contribution to the data record, and peaks close to the frequency where the spectral shape is located. Basically, the filter-bank, instead of being tuned to a single carrier, is tuned to the spectral shape. The spectral shape used to tune the filter-bank is refereed as the candidate spectrum. The major motivation for this procedure was the proper spectrum labelling of licensed users and interferers in cognitive radio applications.