宽带压缩感知信号源个数估计算法.pdfVIP

信号处理 Vo l. 29 No.5 第 29 卷第5 期 2013 年5 月 JOURNAL OF SIGNAL PROCESSING May 2013 宽带压缩感知信号源个数估计算法 王静彭华 (信息工程大学信息系统工程学院，河南郑州 45∞02) 摘 要:基于压缩感知的信号采集系统是目前宽带信号采集的重要发展方向，将其应用于通信侦查或频谱监测 等宽带接收任务时，接收信号实际是处于宽带频率范围内的多个时域重叠的窄带信号，此时多分量压缩感知信 号中的信源个数就成为后续信号处理的关键。本文针对调制信号在循环平稳域具有稀疏性和可区分性，提出一 种基于重构稀疏循环谱特征的信源个数估计算法。算法以重构误差的变化率作为迭代终止条件，应用改进的正 交匹配追踪算法重构;昆合信号的循环谱，并应用循环谱的对称性实现信源个数估计。仿真表明，算法在-12dB 条件下， f言源个数估计准确率可以达到99% 。 关键词:压缩感知:循环谱;线性关系;压缩测量值 中图分类号: TN91 1. 7 文献标识码:A 文章编号: 1003-0530( 2013) 05-0632-08 Source Number Estimation Algorithm for Wide-band Compressed Sensing Signal W ANG Jing PENG Hua (Institute of Information System Engineering , Information Engineering University ,Zhengzhou ,Henan ， 450∞2) Abstract : The signal ac甲lÎsition systems based on Compressed Sensing is the major development orientation of wideband signal sampling. 鄂、四 it is used in communication reconnaissance or spectrum sensing , the received signal is a mixing of 50me narrow-b缸ld signals within wide range.ηlen the number of the mixed signals is the key to subsequent signal analysis. The paper pro阳帽d a 50urce number estimation algorithm focus on the sparsity and distinctive features for modulated signal on the cyclic power level . The algorithm aimed at reconstructing the cyclic power spectrum by the improved OMP , and the 50旧Ce number could be estimated with symmet巧. Simulation results show it has a g!X刘 performance of resolution when the SNR is -12dB. Key words: Compressed Sensing; Cyclie Power Spectrum; Linear Relationship; Compressed Measurements 1 引言