Knowledge-Based Approach to Information Fusion for the Support of Military Intelligence.

摘要

Intelligence cells have to process and evaluate current information to produce timely and reliable pictures of the battlespace. The knowledge-based approach to information fusion presented here focuses on the heuristic human evaluation process. Military information processing is modeled as a context- dependent and template-based heuristic reasoning process and the real task is modeled by a closed world representation. This method of information fusion is independent of a specific military scenario. It is based on the assumption that forces are organized in a structured manner and that they operate in a reasonable and typical way. Respective rules and doctrines can be used as matching templates for knowledge-based reasoning. Using this approach, the analysis and fusion of incomplete and imperfect information and background knowledge can be dealt with in an automated system. The paper discusses the military intelligence cycle; the objectives of information fusion in intelligence processing; fundamentals of heuristic intelligence processing; and knowledge-based processing methods, including context-dependent collation, classification, template-based aggregation, generation of aggregation objects, aggregating hypotheses, the fusion step, and hypothesis management. Twenty briefing charts summarize the presentation. (6 figures, 17 refs.).