2017/05/23 by Woongsup Lee, Lee, Woongsup, Minhoe Kim +3 · 6 citations
Computer Science · Engineering · #Distributed Sensor Networks and Detection Algorithms #Cognitive Radio Networks and Spectrum Sensing #Molecular Communication and Nanonetworks
paper · pdf · doi:10.48550/arxiv.1705.08164
In this paper, we investigate cooperative spectrum sensing (CSS) in a\ncognitive radio network (CRN) where multiple secondary users (SUs) cooperate in\norder to detect a primary user (PU) which possibly occupies multiple bands\nsimultaneously. Deep cooperative sensing (DCS), which constitutes the first CSS\nframework based on a convolutional neural network (CNN), is proposed. In DCS,\ninstead of the explicit mathematical modeling of CSS which is hard to compute\nand also hard to use in practice, the strategy for combining the individual\nsensing results of the SUs is learned with a CNN using training sensing\nsamples. Accordingly, an environment-specific CSS which considers both spectral\nand spatial correlation of individual sensing outcomes, is found in an adaptive\nmanner regardless of whether the individual sensing results are quantized or\nnot. Through simulation, we show that the performance of CSS can be improved by\nthe proposed DCS with low complexity even when the number of training samples\nis moderate.\n