A Hybrid System for AI Enhanced Information Processing and Situation Assessment

We propose to develop a hybrid system incorporating expert systems, Bayesian belief networks, and case-based reasoning to support battlefield multisensor data fusion for situation assessment (SA). Expert systems will provide a framework for the system and give it the capacity to emulate the expert human reasoning heuristics, coordinate overall SA functions, and support numerical and symbolic processing algorithms. Bayesian belief networks will be used to incorporate uncertain data from multiple (and possibly conflicting) sources into the situation assessment, and provide the framework for computing the most likely hypothesis for explaining and making sense of a set of observed tactical situation templates and libraries to be used for interpreting new battlefield activities, events, and observations. Our proposed Phase I study will focus on demonstration and feasibility analysis of a hybrid system for battleifeld situation assessment. We propose a five task effort: 1) determine a tactical scenario; 2) formulate and design a prototype hybrid situation assessment system; 3) develop and implement prototype; 4) evaluate performance; and 5) make recommendations for full-scope development and validation.