Cognitive wideband spectrum sensing using cosine-modulated filter banks

Abstract

A multichannel joint spectrum sensing strategy based on cosine-modulated filter banks (CMFBs) was developed to improve sensing efficiency. The received wideband signal was split into several bands through the filters that are constructed by grouping continuous sub-band filters. Through flexibly designing prototype filter, not only the spectrum of non-uniform bandwidth can be estimated, but also the spectral leakage between adjacent channels can be adjusted. The probabilities of false alarm and detection for multichannel jointly spectrum sensing in the Rayleigh fading channel were deduced. The decision thresholds of different channels were obtained as regards the probability of false alarm. Simulation results show that compared with the traditional energy detector, the detection capability and sensing efficiency have been improved, especially at low signal-to-noise ratio. The CMFB-based multichannel joint sensing scheme not only increases the efficiency of detection, but also enhances the flexibility on the control of bandwidth and spectral leakage between neighbouring channels.