Spectrum sensing is one among the major functions of Cognitive radio systems. Meanwhile, wideband spectrum sensing is a challenging issue with these systems. Conventional wideband techniques require analog-to-digital converters operating at Nyquist sampling rates. Sub-Nyquist sampling or compressed sensing techniques, however, require low rate analog-to-digital converters. In order to avoid interference to primary user transmissions, the cognitive users should perform continuous spectrum sensing and identify the active primary bands even at low-signal-to-noise ratios. In this paper, we propose a sub-Nyquist wideband spectrum sensing technique using multicoset sampling and revised orthogonal matching pursuit (OMP). Conventional orthogonal matching pursuit algorithm uses n iterations to recover n-sparse signal. Extending the iterations beyond n further improves the performance, for instance, extended OMP (OMPa) and sparsity unaware OMP (OMP oo) achieves a better detection capability by increasing the number of iterations beyond n. We analyse the performance of multicoset sampling based wideband spectrum sensing with various OMP algorithms.
展开▼
