Name
Abstract | ||
|---|---|---|
An autonomous agent equipped with sensors is expected to efficiently perform a given mission within a dynamic and changing world of which it has an uncertain representation. In order t o keep this representation well-grounded, the agent has to organi ze the ac- quisition of information autonomously. A key issue is to choose a framework for the representation of beliefs which is adapte d to deal both with time and uncertainty. A perception strategy, desi gned for the purpose of collecting the right information at the right time, should be based on a dynamic evaluation of the relevance of the beliefs of the system. A review of works concerning the problems of belief revision and temporal representation of uncertainty is pre sented. The differences between the variety of existing formalisms are discussed and the framework of a well-adapted approach is recalled. The issue of a robust perception strategy is addressed within this fra mework ; a first solution, together with the corresponding results of simulations of an autonomous surveillance system, are presented. The issue which is dealt with refers to an autonomous system hav- ing an uncertain representation of the changing environment within which it has to fulfil a given mission. The paper addresses theproblem of how the system is taking time into account within its representa- tion of beliefs about the world. An overview of some approaches in this field of research is first presented in order to shed a new light on the proposed approach, which extends earlier works on the representation of uncertainty. The emphasis is laid on the t emporal erosion of the beliefs and on the design of a perception strat egy for an autonomous surveillance system in a changing environment. The proposed approach was tested with simulations of a Markovian model of the evolution of the environment : the results seem very promising for further developments. |
Year | Venue | Keywords |
|---|---|---|
1996 | ECAI | autonomous agent,belief revision |
Field | DocType | Citations |
Computer science,Artificial intelligence,Perception | Conference | 3 |
PageRank | References | Authors |
0.51 | 16 | 1 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Patrick Fabiani | 1 | 40 | 6.67 |