首页 | 本学科首页   官方微博 | 高级检索  
     检索      

自适应步长EASI算法研究及改进
引用本文:余华,吴文全,刘忠.自适应步长EASI算法研究及改进[J].舰船电子工程,2009,29(10):118-121.
作者姓名:余华  吴文全  刘忠
作者单位:海军工程大学电子工程学院,武汉,430033
基金项目:国家"八六三"项目"基于多智能体的传感器网络协同目标跟踪技术研究" 
摘    要:EASI算法(一种在线LMS)是很重要的盲信号分离(BSS)算法,总结了EASI算法应用于盲信号分离的优点与不足,并提出了利用模拟退火策略有效地结合两种步长的优点,在初期快速下降,在后期精确收敛。结果表明,在该实验中使用模拟退火策略来调整学习率简单有效。

关 键 词:盲信号分离  EASI算法  串音误差(ETC)  模拟退火策略

Research and Improvement of Blind Source Separation Based on EASI Algorithm
Yu Hua,Wu Wenquan,Liu Zhong.Research and Improvement of Blind Source Separation Based on EASI Algorithm[J].Ship Electronic Engineering,2009,29(10):118-121.
Authors:Yu Hua  Wu Wenquan  Liu Zhong
Institution:Yu Hua Wu Wenquan Liu Zhong (Dept. of Command Automation College of Electrical Engineering, Naval Univ. of Engineering, Wuhan 430033)
Abstract:EASI agorithm is very important in the Blind Signal Seperation(BSS) area, which is a kind of on-line LMS agorithrn. EASI agorithm is further studied using cross talking error as an evaluation parameter. Smulation result and performance comparison of EASI agorithm is given. Validity of EASI agorithm is proved and the advantages and shortcomings of it is discussed, then a new method which using genetic algorithm to adjust step-length is presented. The method integrates the advantages of large step-length and short step-length. Simulation result of the method shows it is simple and effective.
Keywords:blind signal separation  EASI algorithm  cross talking error(ETC)  simulation analysis
本文献已被 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号