Dynamic Multi-User Intelligence Fusion (DMIF)
As a key contributor to the DARPA DMIF program, IET developed a fundamental innovation in military intelligence modeling and fusion that provides the flexibility for situation adaptation, both for automatic data fusion functionality and interactive human-machine intelligence analytical tasks. Scientists at IET discovered that based on data extracted directly from multi-source imagery as well as DIA sources, dynamic missile order of battle (MOB) for multiple Middle Eastern armies (Iraq, Iran and Syria) could be estimated with a single set of "generic" MOB model fragments. Customers included DARPA and the DMIF system integrator Lockheed Martin. Elements of DMIF technology were transitioned to the U.S. Army's All Source Analysis System PMO in 1998.
Activities on DMIF include: 1) IET's schemas for representing military situations, including representations of uncertainty and models of source credibility, 2) agile modeling tools to generate situation-specific models from network fragments and to fuse evidence from multiple sources, and 3) tools to move robustly among representations, including Bayesian networks, frames and logic.