Spectral Overlap Based Compressed Spectrum Sensing for a Wireless Mesh Network

Detection of the spectrum holes in a cognitive radio network must be carried out very fast with minimum number of input samples. In this paper, a spectrum sensing algorithm has been proposed using compressed sensing technique which requires fewer samples than predicted by Nyquist sampling provided that the licensed user's signal must be sparse in frequency domain. An ambiguity in determining the channel occupancy by compressed sensing method can be overcome by calculating spectral overlap factor from different channels in a wireless mesh network scenario. Sometimes it is misleading whether the channel is occupied by itself or the leakage power from the neighbouring channels which is solved and shown by way of an example.