Novel Frequency Domain Cyclic Prefix Autocorrelation Based Compressive Spectrum Sensing for Cognitive Radio

Abstract

Cognitive radio (CR) has received increasing attention and is considered an important solution to the spectral crowding problem. The main idea behind CR technology is to utilize the unused spectral resources which are determined to be available for secondary user by effective spectrum sensing techniques. However, CR technology significantly depends on the spectrum sensing techniques which are applied to detect the presence of primary user (PU) signals. This paper focuses on detecting OFDM primaries using novel frequency-domain cyclic prefix (CP) autocorrelation based compressive spectrum sensing algorithms. To counteract the practical wireless channel effects, frequency domain approaches for PU signal detection are developed. The proposed spectrum sensing method eliminates the effects of both noise uncertainty and frequency selective channels. Using the frequency domain autocorrelation approach results in highly increased flexibility, facilitating robust wideband multi-mode, multi-channel sensing with low complexity. It also allows to sense weak PU signals which are partly overlapped by other strong PU or CR transmissions.