Integration of all data fusion levels using a blackboard architecture
conference paper
The data fusion model generally accepted, with its process refinement capability, implicitly supposes that all levels of fusion are integrated. A tight integration is essential to gain the maximum benefits from the data fusion process. Unfortunately, the research and development effort in the data fusion domain has generally been done in a fragmented way. Most of the time, the functionalities of one level of fusion have been studied independently from the other levels and, consequently, they have also been implemented on independent test beds. This paper proposes a framework for integrated data fusion composed of a number of interconnected data fusion sub-processes or agents within a closed loop environment. All data fusion agents within the framework are required to comply with the data fusion notion of process refinement. The paper also presents an approach, making use of a blackboard architecture, to achieve the required integration of all data fusion levels. Blackboard systems are well known for their inherent potential to serve as the integration framework for problem-solving systems of a collaborative type. Finally, the migration and expansion of an existing sensor fusion test bed towards a new proof of concept tool for sensor fusion and situation awareness, based on a commercial blackboard system architecture, is discussed.
TNO Identifier
294100
Source title
Proceedings 1998 International Conference on Data Fusion, EuroFusion'98, October 1998, Great Malvern, UK
Collation
9 p.
Pages
195-202
Files
To receive the publication files, please send an e-mail request to TNO Repository.