ITR: Collaborative Research: Distributed Information Fusion Networks for Threat Detection and Assessment

University Of Notre Dame, Notre Dame IN

Investigators

Abstract

Information available for security threat detection/assessment is often generated from distributed heterogeneous sources. Although it may possess a significant qualitative component and lack sufficient time synchrony, this data may provide potentially critical information. This work describes a framework for effective knowledge discovery in such environments via the following tasks: refinement and generalization of the Dempster-Shafer belief theoretic framework; development of the data mining methods of association rule mining and classification; and development of methods enabling the use of timing information in the fusion process to improve threat detection/assessment tasks. The proposed notions will be validated via the development of an experimental platform consisting of a small-scale fusion and decision-making network located at each of the participating institutions. It will possess various sensing capabilities and will mimic the security zones plus gateways structure that characterizes a typical threat detection/assessment environment. The work significantly advances the state-of-the-art in distributed information fusion environments that forms the basis for numerous other application scenarios. The research results and experimental platforms developed will be integrated into courses that already exist and those to be newly developed; various other outreach activities will also be undertaken.