Sub-sampling Techniques for Sampling a Sparse Wideband Signal in Wireless Communications

Abstract

Background/Objectives: The popularity of Wireless technologies, for applications that urges high bandwidth, increases the demand of Radio frequency (RF) spectrum which leads to scarsity in the spectrum resources. Hence it is required to sense the spectrum of a wideband signal at low sampling rates compared to the Nyquist rate. Methods/Statistical analysis: If the wide band signal is Sparse in frequency domain, then it is used to get over the problem of high sampling rates under the compressive sensing framework. Two different efficient schemes called Modulated Wideband Converter (MWC) and Asynchronous Multi-rate (AMR) are proposed here for wideband spectrum sensing at sub-Nyquist rates. The key features of these methods are also compared with the traditional sub-Nyquist sampling approach called Multi-Coset (MC). Findings: The MC sampling is capable of sampling at different time offsets, but unable to poses channel synchronisation. In MWC sampling, the sparse multiband signal is multiplied by a periodic waveform to shift the spectral components of the band of interest to the origin, then filtered and sampled at low rate. This also makes the computational time to reduce greatly. In AMR sampling, the sparse signal is under sampled at different sampling frequencies to produce sample sets with different aliasing spectra. Later these spectra are compared on a common frequency grid to detect the active bands. This mechanism takes fewer samples there by reduces thee computational cost. The simulation results exhibit these advantages over the traditional sampling techniques for wide band spectrum sensing. Application/Improvements: The trade-off between the selection of MWC for less computational time and MR for low complexity in design makes these sub Nyquist schemes suitable for wide band spectrum sensing in cognitive radio like applications.