Contextual models of Information Fusion
Abstract
The fusion of human-based and physics-based information sources will enhance situationawareness for various tasks such as detection, tracking, classification, prediction, pattern oflife or finding multi-entity relationships. However, the information to be fused has differentcharacteristics, in particular, the nature of the uncertainty is different. The core researchobjective of this project is therefore to determine how observations with different types ofassociated uncertainty can be combined such that: 1) uncertainty is suitably assignednumerically to evidence, and 2) the level of trust in or reliance on a given data source isadequately represented. This project rests on the assumption that the conflicting andambiguous nature of data being analysed during information fusion will lead to contextualeffects in relation to decisions of trust. Therefore, this project will 1) diagnose and analysethe prevalence of contextuality in information fusion with regard to decisions of trust and 2)develop models of contextually sensitive information fusion and express them as probabilisticprograms for operational deployment and testing.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 11, 2017
- Source ID
- FA23861714016
Entities
People
- Peter David Bruza
Organizations
- Air Force Office of Scientific Research
- Queensland University of Technology
- United States Air Force