Wideband Spectrum Sensing using Sub-Nyquist Sampling Approaches

Abstract

Cognitive Radio (CR) is expected to play an important role in 5G wireless communications to meet the challenging requirements of massive Machine-to-Machine (M2M) connectivity and Internet of Things (IoT). CR networks should be capable of wideband spectrum sensing (WSS) to provide opportunistic spectrum access and to abate spectrum scarcity. However, WSS approaches are severely limited by the analog to digital converter (ADC) speeds. Sub-Nyquist samplers alleviate the burden on ADC by compressively sampling a wideband signal. This paper focuses on two such samplers, Analog to Information Converter (AIC) and Modulated Wideband Converter (MWC), both of which exploit spectral sparsity. Specifically, a partitioned WSS scheme is proposed with modified AIC and MWC samplers. Real-time signal in the UHF TV band (470-790) MHz is acquired by a Software Defined Radio (SDR) and occupied/vacant bands are detected using the proposed WSS approach. Orthogonal Matching Pursuit (OMP) and Sparse Bayesian Learning (SBL) based sparse recovery approaches aided in this detection. Support recovery performance from both simulations and experimental investigations show that the proposed approach has a good potential for WSS in M2M and IoT applications.