Affordable Cyclostationarity-Based Spectrum Sensing for Cognitive Radio With Smart Antennas

Cognitive radio (CR) technology is a promising way to improve the bandwidth efficiency of underutilized radio spectra. For practical CR systems with limited a priori knowledge of the primary users' signal characteristics, spectrum sensing is mainly based on energy detection and cyclostationary feature detection. Energy detection is simple and practical but becomes ineffective at a low signal-to-interference-and-noise ratio (SINR). Conventional cyclostationary feature detection based on cyclic spectrum estimation can robustly detect weak signals from primary users by only exploiting the cyclostationarity property of communication signals. However, the high implementation complexity it requires limits its widespread usage. In the literature, the use of smart-antenna technology is suggested to track the locations of the primary users and apply transmit beamforming to avoid spatial interference with their signals. The objective of this paper is to establish adaptive cyclostationary (receive) beamforming as an effective spectrum-sensing method with affordable complexity for multiple-antenna cognitive radio. Specifically, we introduce a new spectrum-sensing method that exploits a recently proposed beamforming algorithm, called the adaptive cross-self-coherent-restoral (ACS) algorithm. The complexity of the resultant algorithm is higher than that of the energy detector but is at least an order of magnitude smaller than that of the previous cyclostationary feature detectors, such as the Fourier spectrum cyclic density analysis method and its multitaper-Loeve version. Their performances for spectrum sensing are empirically evaluated and compared in detail in an example