基于混合型专家系统的目标综合识别模型  

A Comprehensive Target Recognition Model Based on the Mixed Expert Systems

作　　者：张晓刚[1] 刘进忙[1] 刘昌云[1] 

机构地区：[1]空军工程大学导弹学院,陕西三原713800

出　　处：《系统工程与电子技术》2002年第11期108-111,共4页Systems Engineering and Electronics

基　　金：国家高等学校骨干教师资助计划首批资助课题 (GG -810 90 0 3 0 -10 0 0 3 )

摘　　要：鉴于目标识别在防空信息处理中的重要性 ,在对目标信息进行综合利用的基础上 ,利用人工智能原理 ,研究把专家系统、神经网络和证据推理相结合用于目标识别 ,建立了目标综合识别的专家系统模型 ,模拟专家识别思维。给出了知识库的构造方法及推理机制。该模型通过人工神经网络 ,实现了知识的并行处理和自动获取 ,具有自学习功能 ;再学习机制可对系统进行及时调整 ,增加或修改知识库的知识 ,具有很好的人机界面和实用性 ,对目标的综合应用识别具有很好的借鉴作用。Considering the importance of target recognition in the air defense information processing and in the light of the theory of artificial intelligence, a comprehensive target recognition model based on mixed expert systems by combining expert systems(ES), neural networks (NN) and Dempster Shafer(D S) theory is built after the target information is synthesized. The methods for constructing a knowledge base(KB)and its reasoning mechanism are presented. By NN, this model achieves the parallel processing and automatic knowledge acquisition with self learning abillty. Its re self learning mechanism can regulate the whole system in time, add or modify the KB's knowledge. Besides, the model has a fine man computer interface and real time characteristics, and has a great value for the comprehensive target recognition.\;

关 键 词：专家系统 目标识别 防空系统 神经网络 

分 类 号：E917[军事] TP18[自动化与计算机技术—控制理论与控制工程]