东北大学学报(自然科学版) ›› 2003, Vol. 24 ›› Issue (8): 743-746.DOI: -

• 论著 • 上一篇    下一篇

一种信号数目估计的小波包分解优化算法

汪晋宽;薛延波;刘志刚;刘福来   

  1. 东北大学信息科学与工程学院;东北大学信息科学与工程学院;东北大学信息科学与工程学院;东北大学信息科学与工程学院 辽宁沈阳 110004
  • 收稿日期:2013-06-24 修回日期:2013-06-24 发布日期:2013-06-24
  • 通讯作者: Xue, Y.-B.
  • 作者简介:-
  • 基金资助:
    教育部科学技术研究重点项目(02085)

Optimizing algorithm in determining the number of signals using wavelet packet decomposition

Wang, Jin-Kuan (1); Xue, Yan-Bo (1); Liu, Zhi-Gang (1); Liu, Fu-Lai (1)   

  1. (1) Sch. of Info. Sci. and Eng., Northeastern Univ., Shenyang 110004, China
  • Received:2013-06-24 Revised:2013-06-24 Published:2013-06-24
  • Contact: Xue, Y.-B.
  • About author:-
  • Supported by:
    -

摘要: 分析了入射信号数目估计的信息理论准则·针对在低信噪比以及数据快照数目较小的情况下,Akaike信息准则(AIC)和最小描述长度MDL都不能正确估计入射信号数目的问题,提出了采用小波包分解方法对MDL准则进行优化的算法,称之为wpMDL算法·考虑子带分解具有预测偏差小和模式隔离等特点,采用基于一维小波包分解的"伪二维"子带分解算法,不仅有效地改善了MDL准则的估计性能,而且计算量比二维分解降低了·仿真结果验证了提案算法的有效性,无论是对信噪比的变化还是数据快照数目的变化,wpMDL算法都能正确地估计入射信号的数目·

关键词: 信号数目, Akaike信息准则, 最小描述长度, 小波包, 子带分解

Abstract: Detection of number of signals by information theoretic criteria is given. In the occasions of low signal-to-noise ratio (SNR) and large snapshots, incorrect estimation of signal number is often yielded with both Akaike Information Criteria (AIC) and Minimum Description Length (MDL). A wavelet packet decomposition (WPD) based algorithm is proposed to optimize MDL, which is addressed wpMDL. With the features of minimum prediction error and mode isolation, pseudo 2-D subband decomposition algorithm based on one-dimension (1-D) wavelet packet decomposition is introduced to not only improve the estimate performance but also save the computation by O(N4) compared with 2-D WPD.

中图分类号: