Jump to content

EU Centre of ExcellenceISO 9001

ERCIMW3C MemberFraunhofer Project Center

Finding Focus of Interest in freely configured sensor neworks

Finding Focus of Interest in freely configured sensor neworks
Department: Machine Perception Research Laboratory
Start date: 2013. 02. 01.
End date: 2016. 01. 31.
External identifier: OTKA 106374

Project manager

Tamás Szirányi
Tamás Szirányi
Address: 1111 Budapest, Kende u. 13-17.
Room number: K 414
Phone: +36 1 279 6106
Fax: +36 1 279 6292
E-mail: sziranyi.tamasEZT_TOROLJE_KI@EZT_TOROLJE_KIsztaki.hu
Homepage: http://www.sztaki.hu/~sziranyi/


[More information]
[More information]
[More information]
[More information]
[More information]


The project aims to show that finding cooperative structures in dense sensor networks may substitute conventional robot based exploration. This hypothesis follows the fact that in practice, the number of surrounding networked sensors increases very quickly.
These networks might comprise any number of virtual cores of the whole sensor set targeted towards some temporal focus of interest without any external (human) control. With these developments come new challenges, e.g. finding the coherence among different modalities and arbitrary distributions.
We propose a new solution for the exploration of autonomous sensor networks: instead of controlled robot agents, virtual cooperation is explored in large sensor networks to form temporary cooperation. Novel results are expected regarding the virtual cooperation among sensor swarm agents, in the topic of finding cores of attention in arbitrary sensor nets; also, new graph structures for describing geometrical directional connectedness of sensors will be researched. The exploration of possible sensor fields of interests (FOIs) can be done by finding hidden cooperated sensor structures where agents are neither able to identify each member in the group, nor are controlled by a leader.
The measured environment may change in time, and a different viewpoint of a given sensor may produce a set of features that are different from others originating from other arrangements, even if using the same modality.
Our main goal is to gain additional (semantic) information from sensor networks, characterizing their geometrical and statistical relations with the goal of increasing the effectiveness of their collaboration.