Skip to Main content Skip to Navigation
Theses

Evaluation de situations multi-agents à partir d’observations partielles et hétérogènes

Abstract : In this thesis, we consider the problem of situation awareness, and more specifically, in crisis situation. In the crisis situation field, we are interested primarily in threat assessment in a topographical environment with a graph representation.Situation awareness systems propose mostly a passive acquisition of information and they face two major problematics. Firstly, the heterogeneity of information which describes the environment brings a knowledge representation problem. The second problem is a high temporal complexity due to the large amount of information while this dimension is crucial in crisis situation assessment.We propose an automated situation awareness system (ASAAP system) which tries to mitigate these problems. We present a dynamic environment modelling allowing an analysis of most valuable variables for situation assessment in order to reduce the field of possibilities and maximise the information gain. In the crisis situation field, we also propose to apply our system on a more specific domain, the threat assessment. This contribution allows to define and analyse the threat level of each exposed zones with the capacity of understanding the enemy's strategy by both representing its targets and the path to reach them. Finally, we present a preliminary approach about the optimisation of the coverage of most valuable variables by the sensors to maximise the information gain.We generated scenarios inspired from real situations to evaluate ours approaches in maritime and military field. The ASAAP system results show an improvement in situation awareness in its complexity and by its capacity to describes the enemy's strategy in reasonable time.
Document type :
Theses
Complete list of metadata

https://tel.archives-ouvertes.fr/tel-03103332
Contributor : Abes Star :  Contact
Submitted on : Friday, January 8, 2021 - 9:30:09 AM
Last modification on : Saturday, January 9, 2021 - 3:34:39 AM
Long-term archiving on: : Friday, April 9, 2021 - 6:22:52 PM

File

sygal_fusion_28354-vasnier-kil...
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-03103332, version 1

Citation

Kilian Vasnier. Evaluation de situations multi-agents à partir d’observations partielles et hétérogènes. Système multi-agents [cs.MA]. Normandie Université, 2020. Français. ⟨NNT : 2020NORMC207⟩. ⟨tel-03103332⟩

Share

Metrics

Record views

71

Files downloads

23