Abstract

Spectrum sensing is one of the predominant function of cognitive radio technology where the requirement of a\nhigh sampling rate in the sensing of a wideband signal is a challenging issue. In fact, the main problem associated with such\nwideband spectrum sensing process is that it is either impractical or too expensive to exhibit Nyquist sampling on such signal\nbecause of need of complex Analog to Digital converters. Thus, considering these factors, a spectrum sensing scheme using\nmulticoset based sub-Nyquist sampling paradigm has been developed in this research paper. The prime objective of this work\nis to develop a scheme for spectrum sensing in cognitive radio without using more analog to digital converter or RF front end\nconverters. Unlike other conventional approaches of spectrum sensing, where initially the wideband signals are regenerated\nfrom the sub-Nyquist samples, and then the power estimation takes place, here in this developed paradigm, the power\nspectrum of the wideband signal has been sensed directly using statistical characteristics. This has strengthened it for saving\nhuge sampling rates and time for cognitive radio sensing having multiple sub-bands. The enhanced multicoset based\nsampling and later the spectrum estimation and energy detection of the cosets signals using greedy algorithm has\nstrengthened the proposed system to deliver optimal results in terms of universal employment, minimal cosets, location of\nactive bands etc.