认知无线传感器网络中频谱盲检测技术研究

发布时间：2018-05-19 13:28

本文选题：认知无线电 + 认知无线传感器网络　；参考：《南京邮电大学》2017年硕士论文

【摘要】：无线传感器网络(Wireless Sensor Networks,WSN)因应用广泛而备受研究人员关注,但是频谱资源紧缺问题限制了它的发展。将认知无线电(Cognitive Radio,CR)技术应用到无线传感器网络中组成认知无线传感器网络(Cognitive Radio Sensor Networks,CRSN)为解决这一问题提供了方法。为了在不干扰主用户(Primary User,PU)的正常通信的前提下接入授权频谱,认知无线传感器网络的认知节点需要不断地检测周围的频谱资源是否被主用户占用。因此,频谱感知技术不仅是认知无线电技术的重要基础,也是区分认知无线传感器网络与无线传感器网络的重要功能。然而,认知无线传感器网络中的频谱感知检测技术研究仍然面临着一些挑战:首先,现有的采用压缩感知(Compressive Sensing,CS)理论的算法大都需要重构原信号,算法计算复杂度很高;其次,传统的能量检测(Conventional Energy Detection,CED)频谱感知方法需要大量的采样样本数并且在低信噪比时检测性能容易因噪声波动而产生影响。为了解决以上问题,本文提出了两种频谱检测算法:(1)提出一种基于高阶统计量(High-Order Statistic,HOS)的压缩宽带频谱盲检测算法(简称HOS-CWSBD),该算法利用了压缩观测数据使采样数据量大大减少;并采用高阶统计量作为频谱检测的判决量,不需要重构出原信号,计算复杂度降低,在不知道主用户先验知识的情况下,也有良好的检测性能。(2)提出一种小样本能量检测中的双门限协作频谱感知方法,该算法不需要知道主用户的先验知识,并采用双门限减少了低信噪比时认知用户对主用户的干扰,利用多维高斯(Cubeof-Gaussian,CoG)近似处理检测结果,克服了传统能量检测方法因需要足够大的样本数使得传输数据大而导致认知无线传感器网络的节点能量消耗过大的问题,在融合中心(Fusion Center,FC)使用“大多数投票”原则做出最终判决,提高整个系统的检测性能。
[Abstract]:Wireless Sensor Networks (WSNs) has attracted the attention of researchers because of its wide application, but its development is limited by the shortage of spectrum resources. The Cognitive Radio Sensor Networks (CRSNs) are applied to the cognitive wireless sensor networks (WSN) to solve this problem. In order to access the authorized spectrum without interfering with the primary user's normal communication, cognitive nodes of cognitive wireless sensor networks need to constantly detect whether the spectrum resources around them are occupied by the primary users. Therefore, spectrum sensing technology is not only an important basis of cognitive radio technology, but also an important function to distinguish cognitive wireless sensor networks from wireless sensor networks. However, the research of spectrum sensing detection in cognitive wireless sensor networks still faces some challenges: first, most of the existing algorithms based on compressed sensing theory need to reconstruct the original signal, and the computational complexity of the algorithm is very high. The conventional Energy detector (CE) spectrum sensing method requires a large number of samples, and the detection performance is easily affected by noise fluctuation at low signal-to-noise ratio (SNR). In order to solve the above problems, this paper proposes two spectrum detection algorithms: 1) A blind detection algorithm for compressed broadband spectrum based on high-order statistics (HOS-CWSBDN), which uses compressed observation data to greatly reduce the sample data. Using high-order statistics as the decision quantity of spectrum detection, it is not necessary to reconstruct the original signal, and the computational complexity is reduced, without knowing the priori knowledge of the primary user. There is also a good detection performance. 2) A dual threshold cooperative spectrum sensing method for small sample energy detection is proposed. The algorithm does not need to know the prior knowledge of the primary user. The dual threshold is used to reduce the interference of cognitive users to the primary users at low SNR, and the multi-dimensional Gao Si is used to approximate the detection results of Cubeof-GaussianCoG. It overcomes the problem that the energy consumption of nodes in cognitive wireless sensor networks is too large due to the large number of samples needed in the traditional energy detection methods. The final decision is made using the "majority vote" principle in fusion center FCs to improve the detection performance of the whole system.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN925;TP212.9

【参考文献】