Multiterminal Filtering for Decentralized Detection Systems

Te Sun HAN  Kingo KOBAYASHI  

Publication
IEICE TRANSACTIONS on Communications   Vol.E75-B   No.6   pp.437-444
Publication Date: 1992/06/25
Online ISSN: 
DOI: 
Print ISSN: 0916-8516
Type of Manuscript: INVITED PAPER (Special Issue on Communication Theory)
Category: 
Keyword: 
multiterminal information theory,  multiterminal statistics,  distributed sensors system,  decentralized detection,  correlated Gaussian noise,  transmission rate constraint,  multi-dimensional Karhunen-Loeve expansion,  multiterminal North filter,  the LWR fast algorithm,  

Full Text: PDF(574.5KB)>>
Buy this Article




Summary: 
The optimal coding strategy for signal detection in the correlated gaussian noise is established for the distributed sensors system with essentially zero transmission rate constraint. Specifically, we are able to obtain the same performance as in the situation of no restriction on rate from each sensor terminal to the fusion center. This simple result contrasts with the previous ad hoc studies containing many unnatural assumptions such as the independence of noises contaminating received signal at each sensor. For the design of optimal coder, we can use the classical Levinson-Wiggins-Robinson fast algorithm for block Toeplitz matrix to evaluate the necessary weight vector for the maximum-likelihood detection.