DocumentCode :
2062933
Title :
An adaptive diffusion quaternion LMS algorithm for distributed networks of 3D and 4D vector sensors
Author :
Jahanchahi, Cyrus ; Mandic, Danilo P.
Author_Institution :
Commun. & Signal Process. Res. Group, Imperial Coll. London, London, UK
fYear :
2013
fDate :
9-13 Sept. 2013
Firstpage :
1
Lastpage :
5
Abstract :
A diffusion widely linear quaternion least mean square (D-WLIQLMS) algorithm for the collaborative processing of quaternion signals over distributed networks is proposed. We show that the underlying quaternion division algebra and the widely linear model allow for a unified processing of 3D and 4D data, which can exhibit both circular and noncircular distributions. The analysis shows that the D-WLIQLMS provides a solution that is robust to link and node failures in sensor networks. Simulations on benchmark 4D signals illustrate the advantages offered by the D-WLIQLMS.
Keywords :
least mean squares methods; signal processing; wireless sensor networks; 3D data processing; 3D vector sensors; 4D data processing; 4D vector sensors; D-WLIQLMS algorithm; adaptive diffusion quaternion; collaborative processing; diffusion widely linear quaternion least mean square algorithm; distributed networks; linear model; node failures; noncircular distributions; quaternion division algebra; quaternion signal processing; sensor networks; Abstracts; Algebra; Educational institutions; Least squares approximations; Quaternions; diffusion algorithms; distributed networks; widely linear models;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Signal Processing Conference (EUSIPCO), 2013 Proceedings of the 21st European
Conference_Location :
Marrakech
Type :
conf
Filename :
6811799
Link To Document :
بازگشت