Situational Awareness as a Man-Machine Map Reduce Job

Improving situational awareness and accuracy of decisions in complex missions relying on streaming open-source data requires scalable information extraction and fusion in collaboration between Man and Machine reasoning. SoarTech, with its proven track-record of basic and applied research and transition into actual deployment, will bring forward advanced imagery and text processing technology integrated in a Hadoop-based distribution framework from its academic partners CUBRC and Dr. Corso from the University of Buffalo. The resulting stream of entity and behavior recognition events from large-scale unstructured and uncalibrated raw data from many sources is fused with a continuously adapting agent simulation and extended into probabilistic predictions of alternative futures. Human intelligence and knowledge is effectively integrated throughout the recognition-fusion-prediction process through a task-aware collaboration environment. As equal participants in the ongoing information-fusion process, decision-makers gain a deeper understanding of the current scenario and the impact of decision alternatives, and thus arrive at better decisions. Our proposed objectives for Phase-I include the demonstration of the feasibility of recognition of HADR-mission-relevant entities and behaviors and the feasibility of such recognition and the subsequent fusion/prediction at scale. We also propose to develop initial use-cases, identify target transition partners, and draft key component and system architecture designs.