Real-time Multi-Gigahertz Sub-Nyquist Spectrum Sensing System for mmWave

Abstract

A real-time sub-Nyquist wideband spectrum sensing system for millimeter wave (mmWave) implemented on National Instruments mmWave software-defined radio system is presented. Based on compressed sensing theory and multicoset sampling architecture, the system is capable of achieving real-time spectrum sensing of 3.072 $\textGHz $-bandwidth signal at the centre frequency of 28.5 $\textGHz $. Bayesian sparsity estimation and data decimation are applied to realize robust performance of spectrum reconstruction under dynamic spectrum scenarios and enable real-time processing, respectively. This paper presents and comments on the impact of noise corruption, spectrum sparsity on the recovery performance and evaluates two low-complexity sparse recovery greedy algorithms of interest.