Intelligent Integration of Human Cognition

Human cognition has played a major role in many past programs that we have had experience with. This includes the “Affordable Moving Surface Target Exploitation” (AMSTE) and NetTrack programs. Although each program approached the tracking and fusion problem from a different aspect and with various technologies, they both gave us insight into the role that human cognition plays with practically any attempt to fuse various data sources to improve tracking capability. On Phase I, we began architecting solutions that could apply to the problems we were familiar with, which included high value target tracking and track based feature usage to perform ambiguity resolution. As the original architecture was developed, it became clear that the problem was too specific and that the solution was not extensible enough to handle many other mission objectives that were of just as much, if not more, importance. The original architecture was strictly a collaborative filter that operated on user actions. The evolution of this architecture led us to a second solution that utilized a multi-layered finite state machine with incorporated collaborative filter. This approach could be applied to more mission objectives, but was still limited in the solution space since finite trees were designed. This meant that only those solutions defined within the system could be used by the operators. This limitation led us to the third and final architecture, a generalized approach with no preconceived architecture. This solution utilizes neural networks, and is the method to be researched, developed, and implemented on Phase II.