Feasibility of blind signal separation using simultaneous perturbation method

被引：0

作者：

Hamaguchi, K ^{[1
]}

Tsushio, K ^{[1
]}

Maeda, Y ^{[1
]}

机构：

[1] Kansai Univ, Fac Engn, Dept Elect Engn, Suita, Osaka 5648680, Japan

来源：

SICE 2002: PROCEEDINGS OF THE 41ST SICE ANNUAL CONFERENCE, VOLS 1-5 | 2002年

关键词：

blind signal separation; mutual information; simultaneous perturbation optimization method;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Blind signal separation (BSS) is a technique in which original signals are extracted only-based on observed signals. In this paper. we propose a recursive method it obtain the separating matrix based on the mutual. information. via the simultaneous perturbation optimization method, Feasibility of this method was confirmed by comparison with the natural gradient method.

引用

页码：2132 / 2135

页数：4

共 7 条

[1]

Akiyama T., 2001, Transactions of the Institute of Electrical Engineers of Japan, Part C, V121-C, P782

[2]

AMARI S, 1997, P IEEE INT WORKSH WI, P107

[3]

HORIGUCHI T, 2000, STAT PHYSICS INFORMA

[4]

Maeda Y., 1993, Transactions of the Institute of Electrical Engineers of Japan, Part C, V113-C, P402

[5]

MATSUOKA K, 1998, T FUZZY SOC JAPAN, V10, P394

[6] MULTIVARIATE STOCHASTIC-APPROXIMATION USING A SIMULTANEOUS PERTURBATION GRADIENT APPROXIMATION [J].

SPALL, JC .

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1992, 37 (03) :332-341

[7]

TSUSHIO K, APPL BLIND SIGNAL SE

← 1 →