Abstract: A method that uses multi-Agent to resolve the situation recognition problem in land battle is given. The method uses two kinds Agent, management Agent and entity Agent. These two kinds Agent work together to solve this complex problem. Entity Agent calculates similarity that considers terrain and motive ability of other Agent nearby it, and gives a result of target grouping, management Agent distributes information data and union result of entity Agent. Simulation result shows that this method can solve grouping problems of terrain segmentation by traditional methods.

Key words: multi-Agent, target grouping, situation recognition, road matching, similarity computing

摘要： 针对陆战场的态势识别问题，提出一种用于多Agent的态势识别系统战场目标分群方法。在陆战场目标分群过程中，管理Agent与实体Agent相互协作，以相似度计算作为分群的依据，考虑每个实体Agent的地形、相互间机动性能等因素，提出各自的分群方案，由管理Agent分发数据并合并结果。模拟结果表明，该方法能够解决传统方法在地形分割中的分群错误问题。

关键词: 多Agent, 目标分群, 态势识别, 道路匹配, 相似度计算

CLC Number: 

• TP18