基于噪声不确定性和用户状态改变的能量检测算法研究

发布时间：2018-02-23 22:12

本文关键词： 能量检测算法噪声功率不确定性主用户状态改变最小采样时间数据碰撞概率　出处：《重庆邮电大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着无线通信业务和技术的快速增长，对无线频谱的需求也日益增大，当前，无线频谱的匮乏是阻碍无线通信发展的瓶颈之一。认知无线电（CR）通过伺机地接入授权的空闲频谱以提高频谱利用率，对缓解无线频谱资源的紧缺具有重要意义。由于实现简单和复杂度低的优点，能量检测算法已经在空闲频谱检测中被广泛应用，但其检测性能特别容易受到噪声功率不确定性（NPU）、检测期间用户状态改变和低信噪比（SNR）因素的影响。本文在这方面进行深入探讨，主要研究内容和成果如下：针对能量检测算法受到NPU的严重影响，本文给出了一种新的复杂度较低的NPU区间估计算法，并且从理论上分析了估计的噪声功率对能量检测算法信噪比墙（SNR WALL）恶化影响，得出了SNR WALL恶化性定理。进一步，基于门限修正提出了一种改进的能量检测算法以消除SNR WALL恶化。仿真结果表明，本文算法能较为精确地估计NPU区间，并且验证了SNR WALL恶化性定理的正确性；同时，改进的能量检测算法性能要优于RSA算法，降低了SNR WALL恶化，提高了检测的鲁棒性。本文将能量检测算法和经典分布式认知MAC（DC-MAC）进行结合，通过跨层协作的方式分析得出了多个次用户（SU）同时检测同一空闲频谱造成检测结果不可靠。为在低SNR场景达到目标检测概率，推导出了能量检测算法的最小采样时间（MST）。基于MST，提出了一种改进的DC-MAC（ODC-MAC）协议。ODC-MAC整合了物理层频谱感知策略和MAC层数据传输，通过跨层协作提高SU的数据传输可靠性。仿真表明，本文理论分析和实际仿真结果吻合；同时，相对于DC-MAC，ODC-MAC可以提高数据传输可靠性和吞吐量。针对检测期间主用户（PU）的活动状态变化和低SNR造成能量检测算法检测性能严重下降的现象，本文给出了一种加权（weight-p）能量检测算法。为减少实现复杂性和节约需要的功耗，weight-p能量检测算法的最优权值建模成MST的优化问题，分析得出了最优权值和次优权值。仿真表明，在PU状态改变和低SNR的场景下，本文提出的weight-p能量检测算法可以提高SU的检测性能和降低虚警概率，，并且在获得相同检测性能的前提下可以压缩检测时间。由于用户（PU或者SU）随机到达对频谱检测的性能有很大影响，为解决这个问题，本文提出了一种反馈叠加能量检测方案。通过将检测周期后半部分采样点的瞬时能量值累加到检测周期前半部分采样点的瞬时能量上，在不延长检测时间的基础之上，提高整个检测周期的能量统计值。仿真表明，反馈叠加能量检测方案的检测性能要优于现有文献的检测方法，并且可以降低用户之间数据发生碰撞的概率，从而提高SU的吞吐量。
[Abstract]:With the rapid growth of wireless communication services and technologies, the demand for wireless spectrum is also increasing. The lack of wireless spectrum is one of the bottlenecks to the development of wireless communication. It is of great significance to alleviate the shortage of wireless spectrum resources. Because of the advantages of simple implementation and low complexity, the energy detection algorithm has been widely used in idle spectrum detection. However, its detection performance is especially vulnerable to the influence of noise power uncertainty, user state change and low signal-to-noise ratio (SNR) factors during detection. In this paper, the main research contents and results are as follows:. In view of the serious influence of NPU on the energy detection algorithm, a new low complexity NPU interval estimation algorithm is presented in this paper, and the influence of the estimated noise power on the SNR wall deterioration of the energy detection algorithm is analyzed theoretically. The SNR WALL deterioration theorem is obtained. Furthermore, an improved energy detection algorithm based on threshold correction is proposed to eliminate the SNR WALL deterioration. The simulation results show that the proposed algorithm can estimate the NPU interval more accurately. At the same time, the performance of the improved energy detection algorithm is better than that of the RSA algorithm, which reduces the deterioration of SNR WALL and improves the robustness of the detection. In this paper, the energy detection algorithm is combined with the classical distributed cognitive MACU DC-MAC. By analyzing the way of cross-layer collaboration, it is concluded that multiple secondary users simultaneously detect the same idle spectrum, which results in unreliable detection results. In order to achieve target detection probability in low SNR scene, The minimum sampling time of the energy detection algorithm is derived. Based on MSTs, an improved DC-MAC ODC-MACMAC protocol. ODC-MAC integrates the spectrum sensing strategy of physical layer and the data transmission of MAC layer. The reliability of data transmission of Su is improved by cross-layer cooperation. The theoretical analysis is in agreement with the actual simulation results, and the reliability and throughput of data transmission can be improved compared with DC-MAC ODC-MAC. In view of the change of the active state of the primary user during the detection and the phenomenon that the detection performance of the energy detection algorithm is seriously reduced due to the low SNR, In this paper, a weighted weight-p energy detection algorithm is presented. In order to reduce the complexity and save the power consumption of weight-p energy detection algorithm, the optimal weight value is modeled as an optimization problem of MST, and the optimal weight value and the suboptimal weight value are obtained. In the scenario of pu state change and low SNR, the proposed weight-p energy detection algorithm can improve the detection performance of Su and reduce the false alarm probability, and the detection time can be compressed on the premise of the same detection performance. In order to solve this problem, the random arrival of users pu or SUU has a great impact on the performance of spectrum detection. In this paper, a feedback superposition energy detection scheme is proposed. By adding the instantaneous energy value of the second half of the detection cycle to the instantaneous energy of the first half of the detection cycle, the detection time is not extended. The simulation results show that the detection performance of the feedback superposition energy detection scheme is better than that of the existing literature, and the probability of data collision between users can be reduced. Thus, the throughput of Su is improved.
【学位授予单位】：重庆邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN925

【参考文献】