论文部分内容阅读
大量的已有研究工作表明:多用户合作频谱感知可以明显提升检测性能.然而,当网络中用户数目较多,并且用户所分布的空间范围较大时,所有用户均参与合作将会带来巨大的感知开销(例如感知时间和能量消耗).基于无监督学习技术和共识理论的最新进展,在本文中我们提出了一种全分布式的合作频谱感知方案.在所提方案中,仅通过一跳可达的邻居之间的信息交互,具有潜在最佳检测性能的用户会自组织地聚到一起,这些用户进而利用平均共识协议来进行合作频谱感知,然后将感知结果广播至全网用户.为了进行性能比较,进一步给出了最优软合并的一种分布式实现方案.数值结果表明:所提方案获得了与最优软合并方案相近的检测性能,并明显优于已有等增益合并方案和基于位置信息的方案.同时,相比于最优软合并方案,所提方案可以大幅度降低感知开销,并且不需要关于用户本地信噪比的先验信息.
A great deal of research work has shown that multiuser cooperative spectrum sensing can significantly improve the detection performance.However, when there are many users in the network, and users have a large range of spatial distribution, all users will have a great cooperation (Such as perceived time and energy consumption) .According to the latest development of unsupervised learning technology and consensus theory, we propose a fully distributed cooperative spectrum sensing scheme in this paper.In the proposed scheme, Information exchange between reachable neighbors, users with potentially best detection performance, are self-organizing together. These users then use the average consensus protocol for collaborative spectrum sensing and then broadcast the perceptual results to users across the network. For performance comparison, a distributed implementation scheme of optimal soft combining is given. The numerical results show that the proposed scheme achieves similar detection performance as the optimal soft combining scheme and is obviously superior to the existing equal-gain combining Scheme and location-based scheme.At the same time, compared with the optimal soft-combining scheme, the proposed scheme can significantly reduce the perceived overhead, and without Priori information about the user’s local signal to noise ratio.