摘要
为了全面了解基于特征值类的合作频谱感知算法的性能,通过Matlab仿真实验对该类算法进行仿真,以探究门限值随虚警概率的变化、采样次数和认知用户数对检测性能的影响,并分析针对不同信噪比的检测概率和实际的虚警概率。仿真结果显示,基于特征值类的算法不需预知主用户信号和噪声方差信息,能够克服噪声不确定度的影响,相较于传统的能量检测,有着更加稳健的检测性能;基于最小特征值极限分布的改进算法在低信噪比、采样次数和认知用户数较少时,判决门限更低、检测概率更高。
In order to comprehensively analyse the performance of eigenvalue-based cooperative spectrum sensing algorithms,the detailed simulation and analysis of this kind of algorithms are completed by simulation experiments with Matlab based on review.The threshold varies with probability of false alarm,the number of the sampling and the cognitive user （CU）varies with the performance of detection,as well as probability of detection and actual Pf as a function of the signal to noise ratio （SNR）are explored.Simulation results show that the eigenvalue-based algo-rithms do not need any knowledge of the primary user and noise variance in advance,that they are robust against noise variance uncertainty and that they have better performance than the con-ventional energy detection.Among them,the improved algorithms based on the limit distribution of the smallest eigenvalue have lower decision threshold and higher Pd than the existing algo-rithms with low SNR,fewer sampling and CU.
出处
《西安邮电大学学报》
2014年第5期27-33,共7页
Journal of Xi’an University of Posts and Telecommunications
基金
国家科技重大专项基金资助项目(2012ZX03001025-004)
国家自然科学基金资助项目(61271276
61301091)
陕西省自然科学基金资助项目(2012JQ8011)
陕西省国际合作基金资助项目(2013KW01-03)
工业和信息化部通信软科学基金资助项目(2014R33)
作者简介
弥寅（1986-），男，硕士，助教，从事认知无线电研究。E—mail：miyin0404@163．com
卢光跃（1971-），男，博士，教授，主要从事通信信号处理研究。E-mail：tonylugy@163．com
