A blueprint for higher-level fusion systems

• Published in: Information fusion 2009, Vol.10 (1), p.6-24
• Main Author: Lambert, Dale A.
• Format: Article
• Language: English
• Subjects: Cognitive machines; Data fusion; Higher-level fusion; Higher-level fusion interfaces; Impact assessment; Information fusion; JDL model; Object assessment; Semantic machines; Sensor fusion; Situation assessment; Situation awareness; Social machines
• ISSN: 1566-2535; 1872-6305
• DOI: 10.1016/j.inffus.2008.05.007

This paper contends that demands on the data fusion community are beginning to exceed its historical roots in sensor fusion, by requiring greater development of automated situation and impact assessments and more appropriate integration with humans engaged in fusion activity. The paper offers a seven building block blueprint for the design of higher-level fusion systems. The first building block involves a deconstruction of the JDL model to apply it beyond machine based fusion. The second addresses machine representation for automated situation and impact assessments, while the third examines machine reasoning for automated situation and impact assessments. The fourth building block then reconstructs a unified framework for automated object, situation and impact assessments so as to accommodate both of the previous building blocks and the traditional approach to sensor fusion. Distributed data fusion constitutes the subsequent building block. The automated presentation of automated situation and impact assessments serves as the sixth building block, before the issue of human involvement in higher-level fusion systems is canvassed. Existing implementations of the various building blocks are referenced rather than discussed in any detail. The aim of the paper is to expose the overarching framework for these higher-level fusion systems without recourse to their considerable underlying complexity.