Skip to Main content Skip to Navigation

Une architecture multi-agent pour la détection, la reconnaissance et l'identification de cibles

Fabienne Ealet 1
TIMC - Techniques de l'Ingénierie Médicale et de la Complexité - Informatique, Mathématiques et Applications, Grenoble - UMR 5525
Abstract : Automatic target recognition imposes the integration of a large variety of heterogeneous knowledge. This leads us to design an architecture distributing knowledge and performing parallel and concurrent processing. We chose a multi-agent approach with the implementation of specialized agents based on incrementality, cooperation, distribution, focus of attention and adaptability. The architecture being used is a multi-threaded one. It is build around an administrator and allows the formalisation of communication protocol between the agents. These agents differ in their role, their behavior ans information they handle and generate. At a given moment, many agent coexist in the image. Each one behaves in an autonomous way for reaching the data and working out its own strategy according to the available information. Planning is dealt with locally at the agent level. A priori knowledge is specified in a database shared by all the agents. Acquired information is stored in a model of the world. The system evolves along time , which imposes an incremental belief update strategy modelled by bayesian networks.
Document type :
Complete list of metadatas
Contributor : Thèses Imag <>
Submitted on : Monday, February 16, 2004 - 11:44:58 AM
Last modification on : Thursday, November 19, 2020 - 3:56:08 PM
Long-term archiving on: : Friday, April 2, 2010 - 7:52:10 PM


  • HAL Id : tel-00004678, version 1



Fabienne Ealet. Une architecture multi-agent pour la détection, la reconnaissance et l'identification de cibles. Mathématiques [math]. Université Joseph-Fourier - Grenoble I, 2001. Français. ⟨tel-00004678⟩



Record views


Files downloads