Detection of vacant unoccupied frequency sub-bands by compressed measurements

This paper is focused on the problem of detection of vacant unoccupied frequency sub-bands over analyzed frequency band in the all-digital format, which is topical issue for the cognitive radio systems. We developed two detection algorithms: one is based on classical Bayesian procedure, another applies Compressive Sensing paradigm, and researched their accuracy dependence on signal-to-noise ratio and compression level. Application of developed algorithms provides great reduction of the number of operations performed by digital processor under negligible degeneration of accuracy of corresponding estimations. Both developed algorithms have close characteristics, which negligible depend on the number of detected occupied frequency channels.