Home | english  | Impressum | Sitemap | KIT
Dynamic World Modeling with Prior Knowledge Matching for Autonomous Systems

Technischer Bericht

Links:
Autor:

Andrey Belkin

Quelle:

Technischer Bericht IES-2011-12. In: Jürgen Beyerer, Alexey Pak (Hrsg.), Proceedings of the 2011 Joint Workshop of Fraunhofer IOSB and Institute for Anthropomatics, Vision and Fusion Laboratory, Karlsruher Schriften zur Anthropomatik, Bd. 11, KIT Scientific Publishing, 2011.

Seiten:

175-184

Konferenz:

Joint Workshop of Fraunhofer IOSB and Institute for Anthropomatics, Vision and Fusion Laboratory, Triberg, Germany, 22.-27. Juli 2011

ISBN:

978-3-86644-855-1

ISSN:

1863-6489

Modern autonomous systems are challenged by complex timecritical tasks. In order to perform reactive and pro-active activities, these systems have to employ a robust and efficient memory component, particularly a high-level memory structure. This structure represents a world modeling, containing all known information about the surrounding environment. The world modeling includes dynamic models, describing the current state of the environment and its past states, and prior knowledge, containing predefined expert knowledge about possible classes, entities and relations in the surrounding world. Moreover, such high-level memory structure has to deal with multi-modal information flow, gathering and fusing incoming sensor information into existing description, analyze the models in order to determine its quality, the missing information, and possible conclusions, and provide all information to other sub-systems of the autonomous system.