Integrated Intelligent Decision and Information System

To address the Air Force need for neuro-networking techniques for space-based sensors across C2 systems, Physical Optics Corporation (POC) proposes to develop a new Integrated Intelligent Decision and Information (IIDI) system based on a unique integration of neural networks, evolutionary algorithms, and fuzzy expert system. In Phase I the IIDI system was tested on real MightySat data supplied by AFRL/MITRE. The new IIDI grid-based neural network and evolutionary algorithm improved accuracy by 2 to 3 times, execution time by 15 to 25 times, training time by 6 to 20 times, and false alarm rate by 2 to 3 times compared to state-of-the-art neural network systems. Phase II development involves optimization of: highly scalable modular and cooperative neural networks; a fuzzy evolutionary algorithm for reducing network traffic and improving training time; a fuzzy expert system for highly human-like situation analysis; and a knowledge database built with expert input for accurate analysis. Integration of these four enhancements will produce a Phase II IIDI system prototype that will directly meet the Air Force requirements for global-scale C2 systems for cognitive distributed situation assessment.