Multi-agent Data Fusion Systems: Design and Implementation Issues

摘要

The objective of Data Fusion (DF) task is making decisions on the basis of distributed data sources accessible through Intra- or Internet. These sources contain heterogeneous data that can be represented in various structures (relational, transactional, etc.), can be of different nature (images, signals, numbers, etc), and accuracy, be measured in different scales (Boolean, categorical, real), can contain uncertainty, etc. An objective of a DF system is to combine data from many different sources to make decision, for instance, classification of an object or an object state, a situation assessment, etc. Within DF specific tasks three issues are of the most significance. The first is development of meta-model of distributed data sources; the second is development of meta-model of combining decisions produced on the basis of particular data sources and the third concerns architectural issue. According to our opinion, the first and the second tasks can be solved successfully only if we focus on the thorough development of the distributed application ontology playing a basic role in effective solution of many problems entailed by heterogeneity and distribution of data. In turn, the ontology-related solution significantly correlates to the architecture of the respective software. The above tasks and respective software tool considered from the design and implementation viewpoints are in the focus of the paper.