Spectrum estimation for cognitive radar

One of the critical functions of a cognitive radar is to examine a large frequency spectrum in real time, in order to detect the presence of narrowband emitters, which radiate on the same frequencies as a wideband radar. The final goal of spectrum monitoring is to recognize which frequency channels of the radar spectrum are occupied over a certain time period. In order to analyze the whole wide spectrum in real time, the time signal samples must be collected at a much lower rate than the Nyquist rate corresponding to the highest frequency of the considered spectrum. This paper proposes a novel approach based on the suitable correlation and averaging of a number of parallel asynchronous channels, each of which is heavily undersampled with respect to the Nyquist rate. We demonstrate that the analysis of the large spectrum is possible by applying our approach by using a limited number of asynchronous channels.