TY - JOUR
T1 - Blind Separation of Noncircular Sources Via Approximate Joint Diagonalization of Augmented Charrelation Matrices
AU - Gou, Xiaoming
AU - Liu, Zhiwen
AU - Ma, Jingyan
AU - Xu, Yougen
N1 - Publisher Copyright:
© 2014, Springer Science+Business Media New York.
PY - 2015/2
Y1 - 2015/2
N2 - An augmented charrelation matrix (ACM), which can utilize both the conventional and the conjugate statistical information in the complex domain, is proposed. The ACM additionally makes use of the conjugate Hessian matrix (namely conjugate charrelation matrix) of the observations of noncircular sources. A blind separation scheme built on the approximate joint diagonalization (AJD) principle is introduced, which precedes some numerical examples to demonstrate the improved performance of the ACM-AJD approach compared with some algorithms in the literature.
AB - An augmented charrelation matrix (ACM), which can utilize both the conventional and the conjugate statistical information in the complex domain, is proposed. The ACM additionally makes use of the conjugate Hessian matrix (namely conjugate charrelation matrix) of the observations of noncircular sources. A blind separation scheme built on the approximate joint diagonalization (AJD) principle is introduced, which precedes some numerical examples to demonstrate the improved performance of the ACM-AJD approach compared with some algorithms in the literature.
KW - Blind source separation
KW - Characteristic function
KW - Hessian matrix
KW - Joint diagonalization
KW - Noncircularity
UR - http://www.scopus.com/inward/record.url?scp=84922222914&partnerID=8YFLogxK
U2 - 10.1007/s00034-014-9867-5
DO - 10.1007/s00034-014-9867-5
M3 - Article
AN - SCOPUS:84922222914
SN - 0278-081X
VL - 34
SP - 695
EP - 705
JO - Circuits, Systems, and Signal Processing
JF - Circuits, Systems, and Signal Processing
IS - 2
ER -