This Small Business Innovation Research Phase I project addresses research required to develop intelligent software systems for distributed real-time data fusion that will be capable of dealing with the explosive growth in the amounts and types of information available to intelligent information processing systems in dynamic environments. Today, automated techniques deal primarily with mere correlation of information, with more sophisticated analyses available only through largely manual processing. While this approach may work for certain situations, there is little hope that it can keep pace with highly mobile, dynamic systems, in which the numbers and positions of distributed information sources evolve in real time. Scientific Computing Associates, Inc., proposes to investigate development of systems based on a software architecture (the Process Trellis) that is effective for data fusion and that can be dynamically grouped into ensembles and ungrouped as needed. Such an architecture seems essential for applications like vehicular traffic management, air traffic control and others in which individual components may enter or leave the `world view` with great rapidity and at unexpected times. In terms of implementation, they propose to use distributed processors functioning as a single dynamic parallel computer. In Phase I, the researchers plan to build a prototype ensemble Trellis system and study its performance in order to assess the feasibility of a full-scale implementation. If the Phase I research is successful, it will lay the groundwork for building data fusion systems that will be able to maintain real-time behavior in dynamic environments by automatically scaling the use of processing power as required.
