Decentralized Distributed Bayesian Estimation

0368081 - ÚTIA 2012 CZ cze A - Abstrakt
Dedecius, Kamil - Sečkárová, Vladimíra
Decentralized Distributed Bayesian Estimation.
Abstracts of Contributions to 7th International Workshop on Data - Algorithms - Decision Making. Praha: ÚTIA AVČR, v.v.i, 2011 - (Janžura, M.; Ivánek, J.). s. 16-16
[7th International Workshop on Data–Algorithms–Decision Making. 27.11.2011-29.11.2011, Mariánská]
Grant CEP: GA ČR 102/08/0567; GA ČR GA102/08/0567
Výzkumný záměr: CEZ:AV0Z10750506
Klíčová slova: estimation * distributed estimation * model
Kód oboru RIV: BB - Aplikovaná statistika, operační výzkum
http://library.utia.cas.cz/separaty/2011/AS/dedecius-decentralized distributed bayesian estimation.pdf

The increasing spatial complexity of modern adaptive wireless and sensor networks and the diversity of its elements call for reliable estimation of several variables of interest. In most regular cases it can be naturally assumed that each participating element carries (to some degree) valid information about these variables and that they could potentially benefit from cooperation and information sharing. We will refer to this modus operandi as the distributed estimation.The basic setting of the decentralized estimation method put the following limitation on the internodal communication: it divides the network into smaller overlapping units consisting of a node and its adjacent neighbours. This node is allowed to exchange data only within its closed neighbourhood. The Bayesian formulation of the problem and its solution abstracts from any particular model case.
Trvalý link: http://hdl.handle.net/11104/0202537