数据融合在雷达目标识别中的应用

数据融合（Data fusin)技术又称多传感器数据融合或分布式传感,即对多类、多源和多平台传感器数据进行组合,提供有关环境综合态势的一种数据处理技术。该技术在防空中首要解决的问题是,根据不同传感器信息源的信息,将飞机,导弹等不同空中目标航迹分类,实现对目标的探测、跟踪和识别。文中列出了其应用实例和用于雷达目标识别的各种数据传感器,分析了利用目标特征的统计或确定的、无知识库或有知识库的、用判定或信号融合的等类数据融合方案,讨论了加权平均法、表决规则法、模糊理论法、神经网络法等数据融合方法,并以“自动驾驶仪”为例,论证了专家系统在数据融合中的作用,由此提出了以专家系统和人工神经网络相结合的数据／信息融合是解决雷达目标识别之捷径的结论。

The Application of Data Fusion in Radar Target Recognition

Data fusion technique is also called multi sensor data fusion or distributed sensing. Data of multi-class, multi-resources and multi-platform sensor is combined with data fusion technique, provided a data processing technique integrated situation of correlative environment. Some of problems in air defence system are chiefly solved with the technique, for instance, different track of air target such as aircraft, missile etc. is classed, the detecting, tracking and recognizing of target is realized based on information of different sensor information resources. The instance of applying data fusion technique and various data sensor used in radar target recognition are given in this paper. Data fusion schemes for statistics or definition, no knowledge base or knowledge base, and decisions or signals fusion etc. are analyzed, in accordance with target characteristics. Data fusion methods that are weighted averages method, voting rules method, fuzzy theory method, nerve network method etc are discussed. Taken "automatic driving device" as example, the action of expert system in data fusion is demonstrated. The conclusion that data/information fusion combining expert system with artificial nerve network is a shortcut way of solving radar target recognition is presented.