Hybrid Statistical Network/Expert System Approach for Advanced Data Fusion

The human brain processes information from several sensors of the body and extracts meaningful information about the environment. Likewise, the ability to fuse multilevel, multifaceted processes dealing with the association, correlation, and combination of data from single or multiple sources is desired to attain state and identity estimates in a timely manner. However, current data fusion techniques beyond level 1 (correlation) lack automation and adaptation to highly mobile, dynamic systems today. Therefore, advanced data fusion technology is needed to discern complex processes such as multiple target tracking, which fuses sensor data and heuristic information to achieve low level and high level assessments.We propose an approach to all source data fusion that can be applied to a wide range of sensor phenomenologies based on AbTech's Statistical Network_ technology and NASA's CLIPS expert system. Our Statistical Network based ModelQuest Expert_ tool tightly couples Statistical Networks for data modeling, a relational database, and Expert Strategies for heuristic information processing. The CLIPS expert system provides management, integration, automation and implementation of dynamic control, alignment, association, assignment, tracking, and heuristic information such as mission operation. The result is a well defined methodology to produce an intelligent and automated agent for all source data fusion.