Sub-Sampling Quantize-and-Forward Schemes for Relay Networks

In this paper, we consider the sensor network, where multiple sensors communicate with a single fusion center via their respective direct links and relays. We propose schemes exploiting both quantize-and- forward (QF) and compressive sensing (CS) in relay networks, aiming to take full use of the diversity to recover the original sparse signals through the collection of a small number of samples. Such sub- sampling QF framework not only inherits the advantage of QF, but also enjoys the merits of reduced sampling rate by exploiting CS technique. The proposed two sub-sampling QF schemes place emphasis on: i) diversity gain in CS domain by employing different projection matrices in different sensors; ii) diversity gain in coding domain due to the correlations, respectively. We assess and compare the performance of the proposed schemes under AWGN channel. Simulation results demonstrate that the proposed schemes achieve excellent recovery performance with reduced communication overhead.