TECSS: Time-Efficient Compressive Spectrum Sensing Based on Structurally Random Matrix in Cognitive Radio Networks

Abstract

As an advanced technology of implementing wideband spectrum sensing and enhancing the ability of secondary users to utilize multichannel diversity in cognitive radio networks, compressive sensing, without requirement of increasing ADC sampling rate, makes use of unique trait of sparse channel occupancy in cognitive radio networks to detect appearance of primary users in wide spectrum. However, current existing research works aim at highly accurate sensing based on Gaussian Random Matrix (GRM) design, but they fail to take time-efficient sensing into consideration, because GRM causes large computing volume and inefficiency, which lowers the capability of compressive sensing to quickly adapt to channel occupancy change rate of primary users and in turn decreases utility of spectrum exploitation for secondary users. In this paper, we design a Structurally Random Matrix (SRM) by combining GRM and Partial Fourier Matrix (PFM) to improve time efficiency of compressive sensing. As SRM possesses the sensing accuracy merit of GRM and the computing efficiency merit of PFM, the proposed compressive sensing scheme TECSS largely improves time efficiency at a cost of minor sensing accuracy. Simulation results reveal that the sensing accuracy of our proposed TECSS is 92.5% in average sense, slightly below that (95%) of compressive sensing schemes based on GRM, but time-efficiency is upgraded by 100%.