Circuits, Systems, and Signal Processing
https://doi.org/10.1007/s00034-019-01303-x
A Modified RLS Algorithm for ICA with Weighted
Orthogonal Constraint
Jianwei E
1
· Jimin Ye
1
Received: 6 March 2019 / Revised: 29 October 2019 / Accepted: 29 October 2019
© Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract
Independent component analysis (ICA), as an important data processing technique, is
widely employed in many areas. The objective of the ICA is to recover independent
components from observed signals. Several algorithms, such as equivariant adaptive
separation via independence algorithm, least-mean-square (LMS)-type algorithms and
recursive least-squares (RLS)-type learning rules, are proposed to solve the ICA prob-
lem. In the present paper, a modified RLS algorithm for ICA with weighted orthogonal
constraint is developed to implement source separation based on the local convergence
analysis of the available algorithm. Comparative experiment results demonstrate that
the proposed algorithm is better than existing learning rules in the aspect of the accu-
racy of separation and stability.
Keywords Independent component analysis · Least-mean-square algorithm ·
Recursive least-squares algorithm · Weighted orthogonal constraint
1 Introduction
The purpose of the ICA is to estimate statistically independent source signals from
the observed signals [10]. Here, the observed signals are linearly mixed from source
signals, and there is nothing about mixing coefficient [8]. Hence, ICA is called blind
source separation (BSS) by the majority of academics.
Since ICA technique was formally introduced in 1994 [5 ], it has been widely applied
in various fields. For instance, in biomedical signal processing, Rejer and Górski
[13] utilized different ICA algorithms to produce the brain computer interfaces, and
Yang and Tavassolian [22] canceled the motion noise of cardio-mechanical signals by
the constrained ICA. In the financial series analysis, ICA was employed to separate
B
Jimin Ye
jmye@mail.xidian.edu.cn
Jianwei E
jianwei_math@163.com
1
School of Mathematics and Statistics, Xidian University, Xi’an 710071, China
评论0
最新资源