计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (11): 142-144.DOI: 10.3778/j.issn.1002-8331.2009.11.043

• 数据库、信号与信息处理 • 上一篇    下一篇

利用禁忌遗传和原子特性实现信号稀疏分解

袁志刚,舒维杰,尹忠科,王建英   

  1. 西南交通大学 信息科学与技术学院,成都 610031
  • 收稿日期:2008-02-26 修回日期:2008-05-12 出版日期:2009-04-11 发布日期:2009-04-11
  • 通讯作者: 袁志刚

Signal sparse decomposition based on TS and GA and atom property

YUAN Zhi-gang,SHU Wei-jie,YING Zhong-ke,WANG Jian-ying   

  1. School of Information Science & Technology,Southwest Jiaotong University,Chengdu 610031,China
  • Received:2008-02-26 Revised:2008-05-12 Online:2009-04-11 Published:2009-04-11
  • Contact: YUAN Zhi-gang

摘要: 阻碍信号稀疏分解运用于信号处理产业化的主要原因,是由于信号的稀疏分解的计算量十分巨大。利用基于Matching Pursuit(MP)方法实现的信号稀疏分解算法,采用遗传算法(GA)和禁忌搜索(TS)相结合,快速寻找MP过程中每一步分解的最佳原子,最后再利用原子的特性进一步的优化。实验结果表明,该算法提高了信号每一步MP分解中寻找最佳原子的能力,并由此提高了信号稀疏分解的速度。

关键词: 信号处理, 稀疏分解, Matching Pursuit(MP)方法, 遗传算法, 禁忌搜索, 原子特性

Abstract: Sparse decomposition hampered signal applied to signal processing in industrial production is mainly due to the sparse signal decomposition very huge amount of computation large.Based on the Matching Pursuit(MP) method signal sparse decomposition algorithm,Genetic Algorithm(GA) and Tabu Search(TS),which combines quick find MP process every step of the best atomic decomposition.Finally,the use of atomic properties further optimization.Experimental results show that the proposed algorithm enhances the signal in each step of decomposition MP for the best possible atomic capabilities,and improve the speed of decomposition sparse.

Key words: signal processing, sparse decomposition, Matching Pursuit(MP), Genetic Algorithms(GA), Tabu Search(TS), Atom Property