一种基于峭度累积量比例微分控制的盲源分离学习率

doi:10.3969/j.issn.0372-2112.2015.05.015

电子学报 ›› 2015, Vol. 43 ›› Issue (5): 929-934.DOI: 10.3969/j.issn.0372-2112.2015.05.015

一种基于峭度累积量比例微分控制的盲源分离学习率

陈国钦

福建师范大学福清分校电子与信息工程学院, 福建福清 350300

收稿日期:2014-07-02 修回日期:2014-10-29 出版日期:2015-05-25
- 作者简介:
- 陈国钦 男,1962年生于福建永春.现为福建师范大学福清分校电子与信息工程学院副教授.主要研究为方向电子、通信和信号处理.E-mail:cgq6203@163.com
- 基金资助:
- 福建省教育厅A类重点项目 (No.JA13341)

A Learning Rate in Blind Source Separation Based on Proportional Differential Control of Kurtosis Cumulative

CHEN Guo-qin

School of Electronic and Information Engineering, Fuqing Branch of Fujian Normal University, Fuqing, Fujian 350300, China

Received:2014-07-02 Revised:2014-10-29 Online:2015-05-25 Published:2015-05-25

摘要/Abstract

摘要：

自然梯度算法由于良好的分离性能在盲源分离中占有重要的地位,但该算法基于固定步长时,无法很好兼顾收敛速度和稳态误差.本文借鉴自动化控制的PID(Proportion Integration Differentiation)算法,提出一种与分离状态紧密结合的变步长学习率算法.由于完成分离的信号峭度累积量是一个固有值,分离过程的信号峭度累积量与固有值将有一个不断减小的误差值.该算法以指数函数值来体现该误差值.再利用该误差构成比例微分的变步长算法,其中的步长初始值就相当于控制误差的比例值,而误差的微分项则得到加速的调整值.该算法仿真实验结果与固定步长自然梯度盲源分离算法的仿真实验结果对比:对应于初始步长的一个最大值和一个最小值,该算法的两次迭代次数均低于采用固定步长算法的迭代次数,并且对于不同类型信号在两次迭代次数间的差值约10～40次,而两种算法的稳态误差是相同的.

关键词: 盲信号分离, 峭度累积, 比例微分控制, 变步长学习率

Abstract:

Natural gradient algorithm occupies an important position in blind source separation due to its good separation performance,but when the algorithm is based on a fixed-step size,a good balance will impossibly be achieved between the convergence rate and steady-state error.This article drew PID (Proportion Integration Differentiation) algorithm of automation control as a reference and proposed an algorithms in variable-step learning rate closely integrated with the state of separation.Due to the fact the cumulative amount of the signal kurtosis was an intrinsic value after the complete separation,there arose a gradually decreasing error value between the cumulative amount of the signal kurtosis of the separation process and the inherent value.The exponential function value of e in the algorithm was applied to reflect the error value.Then the error was used to constitute proportional differential variable-step algorithm,among which the initial value of the step was equivalent to proportional value of the error control,and the differential term of the error gained the adjusted values in acceleration.The simulation results show that corresponding to a maximum and a minimum of step initial value,the number of iterations of the algorithm in two times was lower than that of iterations with fixed-step algorithm,and the difference between the two iterations was about 10 to 40 times for signals of different type,however,the steady-state error of the two algorithms was the same.

Key words: blind signal separation, kurtosis cumulative, proportion differential control, variable-step size in learning rate

中图分类号:

TN911

陈国钦. 一种基于峭度累积量比例微分控制的盲源分离学习率[J]. 电子学报, 2015, 43(5): 929-934.

CHEN Guo-qin. A Learning Rate in Blind Source Separation Based on Proportional Differential Control of Kurtosis Cumulative[J]. Acta Electronica Sinica, 2015, 43(5): 929-934.

参考文献

[1] 孙守宇,郑君里,吴里江,赵莹.峭度自适应学习率的盲信源分离[J].电子学报,2005,33(3):473-476. Sun Shou-yu,Zheng Jun-li,Ww Li-jiang,Zhao Ying.Blind source separation with kurtosis adaptive learning rate[J].Acta Electronica Sinica,2005,33(3):473-476.(in Chinese)
[2] Yang H H,Amari S.Adaptive on-line learning algorithms for blind separation-maximum entropy and minimum mutual information[J].Neural Computations,1997,(9):1457-1482.
[3] Cardoso J F,Laheld B.Equivariant adaptive source separation[J].IEEE Trans Signal Processing,1996,44:3017-3030.
[4] Cruces S,Cichocki A,Castedo L.An iterative inversion approach to blind source separation[J].IEEE Transactions on Neural Networks,2008,11(6):1423-37.
[5] 付绍军,赵国栋,卓琨.一种新的多信号卷积混合信号盲分离算法[J].电讯技术,2012,52(3):328-332. Fu Shao-jun,Zhao Guo-dong,Zhuo Kun.A new blind source separation algorithm for convolved multiple source signals[J].Telecommunication Engineering,2012,52(3):328-332.(in Chinese)
[6] 柯维,张永祥,吕博.基于微分进化算法的盲源分离[J].海军工程大学学报,2012,24(5):12-17. Ke Wei,Zhang Yong-xiang,Lu Bo.Blind source separation based on differential evolution algorithm[J].Journal of Naval University of Engineering,2012,24(5):12-17.(in Chinese)
[7] 张银雪,田学民,邓晓刚.基于改进人工蜂群算法的盲源分离方法[J].电子学报,2012,40(10):2026-2030. Zhang Yin-xue,Tian Xue-min,Deng Xiao-gang.Blind source separation based on modified artificial bee colony algorithm[J].Acta Electronica Sinica,2012,40(10):2026-2030.(in Chinese)
[8] 陈琛,马庆伦,李灯熬,赵菊敏.NGA实现互信息量最小化的盲源分离[J].计算机工程与应用,2013,49(4):121-124. Chen Chen,Ma Qing-lun,Li Deng-ao,et al.NGA achieves blind source separation with mutual information minimum[J].Computer Engineering and Applications,2013,49(4):121-124.(in Chinese)
[9] 刘俊良,于凤芹.基于分离度变步长的自然梯度算法的语音分离[J].计算机工程与应用,2013,49(21):216-219. Liu Jun-liang,Yu Feng-qin.Speech separation based on variable step size natural gradient algorithm depending on sepa-ration degree[J].Computer Engineering and Applications,2013,49(21):216-219.(in Chinese)
[10] 李思怡,王永威,黄琰,陈惠娟.一种基于自然梯度的两步盲源分离算法[J].微电子学与计算机,2013,30(6):169-172. Li Si-yi,Wang Yong-wei,Huang Yan,Chen Hui-juan.Blind sourcr separation based on two-segment natural gradient algorithm[J].Microelectronics & Computer,2013,30(6):169-172.(in Chinese)
[11] 吕淑平,祝捷.一种改进的自适应混合神经网络盲分离算法[J].计算机应用研究,2013,30(4):1055-1057. Lv Shu-ping,Zhu Jie.Improved self-adaptive mixing neural network algorithm for blind source separation[J].Application Research of Computers,2013,30(4):1055-1057.(in Chinese)
[12] Zeng T J,Feng Q Y.A parallel dual matrix method for blind signal separation[J].Neural Computation,2013,26(3):592-610.
[13] 胡剑峰.PID参数继电自整定方法的研究[J].仪表技术,2012,(06):39-41,44. Hu Jian-feng.The PID parameter self-tuning method research[J].Instrumentation Technology,2012,(06):39-41,44.(in Chinese)
[14] 焦嵩鸣,赵潇明,侯晓宁.自整定PID控制算法在优化控制站中的应用[J].计算机仿真,2014,31(1):412-415. Jiao Song-ming Zhao Xiao-ming,Hou Xiao-ning.Self-tuning PID control algorithmin application of optimal control station[J].Computer Simulation,2014,31(1):412-415.(in Chinese)

一种基于峭度累积量比例微分控制的盲源分离学习率

A Learning Rate in Blind Source Separation Based on Proportional Differential Control of Kurtosis Cumulative

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 11

编辑推荐

Metrics

[1]	张登科. 基于ICA的盲信号分离正定性检验方法[J]. 电子学报, 2012, 40(11): 2303-2308.
[2]	张和发;李立萍;杨小牛;李万春. 一种适用于微弱信号盲提取的白化方法[J]. 电子学报, 2011, 39(6): 1297-1301.
[3]	罗志增;曹铭. 基于最大信噪比盲源分离的脑电信号伪迹滤波算法[J]. 电子学报, 2011, 39(12): 2926-2931.
[4]	谢胜利;孙功宪;肖明;傅予力;吕俊. 欠定和非完全稀疏性的盲信号提取[J]. 电子学报, 2010, 38(5): 1028-1031.
[5]	柳艳红;魏学业;吉松坡. 后非线性混合盲信号分离的一种新算法[J]. 电子学报, 2008, 36(7): 1401-1404.
[6]	马丽艳;李宏伟. 一种基于非线性PCA的卷积混合盲源分离算法[J]. 电子学报, 2008, 36(5): 1009-1012.
[7]	何培宇, 周激流, 夏秀渝, 王永德, 赵刚. 一种基于二阶盲信号分离的多路声回波抑制模型[J]. 电子学报, 2006, 34(11): 2109-2114.
[8]	谢胜利;王杰. 一个基于盲信号分离的多路回波消除结构[J]. 电子学报, 2004, 32(7): 1124-1126.
[9]	游荣义;陈忠. 一种基于ICA的盲信号分离快速算法[J]. 电子学报, 2004, 32(4): 669-672.
[10]	冯爱刚;殷勤业;邓科. 基于时空解析恒模算法的CDMA智能天线系统盲信号分离[J]. 电子学报, 2003, 31(3): 328-331.
[11]	何培宇;殷斌;P.C.W.Sommen. 一种有效的语音盲信号分离简化混合模型[J]. 电子学报, 2002, 30(10): 1438-1440.