Wideband Sensing For Sparse Spectrum

As the enormous portions of the functioning spectrum being under-utilized and the growing revolutions of wireless devices,there is a requirement to efficiently apply the scarce frequency spectrum.Cognitive Radios(CRs)or spectrum sharing radios are recognized as one of the clarifications to the spectrum scarcity delinquent.These radios adaptively utilize the unoccupied frequency bands without triggering interference to licensed Primary Users(PUs),and evacuate these bands on sensing the PU activity.To recognize this efficiently with low-latency,these radios should be proficient of sensing a wide spectral range,in the order of a few hundred MHz this needs Wideband Radio-Frequency(RF)front-ends with high-rate Analog-to-Digital Converters(ADCs),followed by digital processing.This naturally expends high power.In case of sparse spectrum,the sampling rates can be reduced considerably depending on its sparse support and the desires on the ADCs can be relaxed.These are frequently formulated as Compressive Sensing(CS)problems,where the conventional detection is performed on the compressive approximation of the signal Unfortunately,recovery of such compressed samples is computationally expensive.We proposed new fast computing DFT method for signal recovery on Wideband sparse spectrum.This method is based on downsampling,and combines MPP with the Bigband method.All operations of solving MPP are linear with analytical solutions involved.The Bigband method is utilized to modify the error of the solution to MPP,caused by the interference of insignificant freq.grids.Theoretical complexity analysis and simulation results demonstrate that the proposed method is hardware-friendly and has a better performance as compared to other reported algorithms.