Spectrum estimation by multiple asynchronous channels

Abstract

The need to estimate in real time a large frequency spectrum, in order to detect the presence of narrowband emitters within a certain frequency area, is relevant in many applications, in particular those related to cognitive radar. The goal of spectrum monitoring is to analyze the whole wide spectrum in real time. This paper starts from previously proposed solutions, based on correlating the undersampled signals of only two asynchronous channels. The drawback of this solution is that, in specific cases, some ambiguities may occur along the estimated spectrum, due to the casual overlapping of some aliased tones, produced by undersampling. This paper demonstrates that, by the use of multiple channels and suitable correlation algorithms, the above ambiguities can be cancelled and the final spectrum estimation comes out to be correct.