基于压缩感知和指数熵的频谱感知技术研究

发布时间：2019-07-03 15:51

【摘要】：随着无线通信业务的发展,可利用的频带资源变得日益紧张。然而在另一方面,很多已被授权的频谱资源的实际利用率是非常低的。认知无线电技术正是针对这一状况提出并发展起来的。在保证授权用户的正常通信不被干扰的前提条件下,通过使认知用户动态地接入授权频段来提高授权频谱的利用率。其中,频谱感知技术是认知无线电得以实现的必要条件。首先对认知无线电技术的提出背景以及相关发展研究状况进行了介绍,然后分别详细讨论了基于发射机以及基于接收机的频谱感知算法的理论模型,并分别对基于匹配滤波器检测、能量检测、协方差矩阵检测以及循环平稳特征检测的频谱感知原理进行了详细阐述,总结了其各自的适用场合及优缺点。然后针对单用户频谱感知的不足,介绍了认知无线电系统多用户合作频谱感知技术,分别对基于硬判决合作和软判决合作的融合原理进行了分析并讨论了其各自优缺点,给出了相应的检测性能分析。针对在宽带频谱感知中按照传统采样方式时数据量较大的问题,提出了一种基于压缩感知和匹配滤波器理论的频谱感知模型。当原始信号的带宽很大且可压缩时,为了节约信道资源,可以先将原始信号进行压缩采样然后再在信道中传输。然而在认知无线电系统中,认知用户可能不具有很高采样速率的能力而传输的信号仍然数据量很大,此时在认知用户处再进行二次压缩采样,然后根据匹配滤波器原理生成判决统计量并进行判决。仿真实验给出了提出算法与传统能量检测算法的检测性能对比,并分别分析了压缩矩阵参数对该算法检测性能与检测时效性的影响。另外,针对目前频谱感知技术对噪声功率不确定性的鲁棒性较差、在低信噪比情况下的检测性能不佳等问题,提出了一种基于指数熵的频谱感知算法。该方法根据H0和H1条件下接收信号频域幅值分布特性的不同,估计接收信号的指数熵,然后通过与预设的门限进行比较,进而判断授权用户信号存在与否。该方法具有不需要信号的先验知识、抗噪声功率不确定性以及在低信噪比下可以得到较高检测概率等优点。最后,受基于能量检测的等增益合并算法的启发,本文还提出了一种基于软判决的指数熵合作检测方案。仿真实验表明,提出的基于指数熵的频谱感知算法对噪声功率不确定性具有鲁棒性,同时,在低信噪比情况下也具有较好的检测性能;提出的基于软判决的多用户合作指数熵频谱感知方案较传统的硬判决合作频谱感知方案也具有更好的检测性能。
[Abstract]:With the development of wireless communication services, the available frequency band resources become increasingly tense. On the other hand, however, many authorized spectral resources have a very low actual utilization. The cognitive radio technology is proposed and developed for this situation. Under the precondition that the normal communication of the authorized user is not disturbed, the utilization rate of the authorized frequency spectrum is improved by dynamically accessing the authorized frequency band by the cognitive user. In which the spectrum sensing technology is a necessary condition for realizing the realization of the cognitive radio. Firstly, the background of the cognitive radio technology and the related development research situation are introduced, then the theoretical model of the frequency spectrum sensing algorithm based on the transmitter and the receiver is discussed in detail, and the detection and the energy detection based on the matched filter are respectively discussed. The spectrum sensing principle of the detection of the covariance matrix and the detection of the smooth characteristic of the loop is described in detail, and their respective application and advantages and disadvantages are summarized. Then, aiming at the shortage of single-user spectrum sensing, the multi-user cooperative spectrum sensing technology of the cognitive radio system is introduced, the fusion principle based on the hard decision cooperation and the soft decision cooperation is analyzed and the respective advantages and disadvantages are discussed, and the corresponding detection performance analysis is given. A spectrum-aware model based on the theory of compression-aware and matched-filter is proposed for the problem of the large amount of data in the wide-band spectrum sensing according to the traditional sampling method. When the bandwidth of the original signal is large and compressible, in order to save the channel resources, the original signal can be compressed and then transmitted in the channel. In the cognitive radio system, however, the signal transmitted by the cognitive user may not have a high sampling rate still has a large amount of data, at which time the secondary compression sampling is performed at the cognitive user, and then the decision statistic is generated and the decision is made according to the matched filter principle. In this paper, the detection performance of the proposed algorithm and the traditional energy detection algorithm is compared, and the influence of the compression matrix parameters on the detection performance and the time-effectiveness of the algorithm is analyzed. In addition, aiming at the problem of poor robustness of the current spectrum sensing technology to the noise power uncertainty, the detection performance under the condition of low signal-to-noise ratio is not good, and the spectrum sensing algorithm based on the index entropy is proposed. The method estimates the index entropy of the received signal according to the difference of the frequency domain amplitude distribution characteristic of the received signal under the condition of H0 and H1, and then judges whether the authorized user signal exists or not by comparing with a preset threshold. The method has the advantages of no prior knowledge of signals, uncertainty of anti-noise power and high detection probability under low signal-to-noise ratio. Finally, an exponential entropy cooperation detection scheme based on soft decision is presented in this paper. The simulation experiment shows that the proposed spectrum-sensing algorithm based on the index entropy is robust to the noise power uncertainty, and also has better detection performance in the case of low signal-to-noise ratio; The proposed soft-decision-based multi-user cooperation index entropy spectrum sensing scheme has better detection performance than the traditional hard-decision cooperative spectrum sensing scheme.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN925

【参考文献】