Research Of Technologies For Situation Assessment In Multi-source Data Fusion

As the level 2 processing of multi-source data fusion, the situation assessment is to distill the situation elements, explain the war field situation, point out the activity patterns, conclude the intentions of enemy and forecast the changes of near feature situation by analyzing the data frames processed by the level 1 fusion.First, the situation awareness technique is discussed. The event detection method and the fuzzy method for events and states are presented, the hierarchy of target classification is discussed and a synthetic algorithm for fulfilling increased group formation and maintaining group structure by using the nearest-neighbor method and field knowledge is proposed. Second, the paper analyzes the characteristics and reasoning mode of situation assessment, and points out that situation assessment is essentially a process of plan recognition and is a gradual solving process to information in different abstract level, the paper also presents that the knowledge is partitioned to four levels: events, actions, missions, and plans, the plan recognition model of situation assessment is given, the implementation of plan recognition algorithms are presented. Third, a template-matching implement technique for achieving situation recognition is studied in depth. CLIPS, a tool of expert system, is used to express events, activities and military plans template.Finally, two scenes are created to test the algorithms for event detection, target classification and plan recognition.