Two-Dimensional Compressive Spectrum Sensing in Collaborative Cognitive Radio Networks

Abstract

Collaborative compressive spectrum sensing techniques are adopted to achieve spectrum reconstruction at sub-Nyquist sampling rate as well as to improve detection reliability in cognitive radio networks. However, most of the existing research of compressive spectrum sensing only focuses on the sparsity in the frequency domain. In pursuit of achieving compressive sampling further beyond Nyquist rate and better recovery performance, this paper presents a two- dimensional (2-D) compressive spectrum sensing scheme in a centralized collaborative setting by applying compression in both frequency and spatial dimensions. To exploit sparsity in spatial dimension, an algorithm based on ant colony optimization aiming at permuting power spectrum observations from users is proposed. Numerical simulations show the superior compressibility and recovery performance of the proposed 2-D compression scheme, as well as effectiveness of the proposed permutation algorithm.