AIT is developing methods for combining multisensor imagery, radar signals, and information into integrated visualizations to aid the analyst. Combining the outputs of our multisensor fused data mining for features and objects, dynamic target grouping and feature learning-aided tracking, provides input to information networks that generate hypotheses about the situation. Visualization of the dynamic scene is interactive and in 3-D, supporting enhanced situational awareness of moving targets.

AIT is pioneering new methods of representing and fusing information. Based on insights into brain dynamics, we are exploiting nonlinear models of integrate and fire neurons that synchronize their spiking to form hypotheses about categorized data in the context of semantic information. Our approach supports the learning of associations among semantic facts, and the learning of concepts, events and trends in hierarchy. Hypotheses emerge as phase-locked groups of nodes spiking in synchrony, and multiple hypotheses are supported as out-of-phase groupings. This approach is being extended to support the inclusion of textual information as well.

*GMTI (ground moving target indicator)