基于神经网络的辐射源识别系统设计  

Radiator identification system based on neural networks

作　　者：张国柱[1] 姜文利[1] 周一宇[1] 

机构地区：[1]国防科技大学电子科学与工程学院,湖南长沙410073

出　　处：《系统工程与电子技术》2004年第2期268-272,共5页Systems Engineering and Electronics

基　　金：国防预研基金资助课题

摘　　要：针对当前辐射源识别系统中存在的问题,提出了一种结合神经网络技术的辐射源识别新方法。该方法可以快速高效的识别各类辐射源,既有基于统计分析的辐射源识别系统的快速性,又有基于专家系统的辐射源识别系统的自适应性和准确性。实际仿真结果表明该方法是有效的,尤其对于参数不全、参数畸变的雷达辐射源,其识别率和识别置信度都有较大提高。在本文方法的基础上,设计出一种结构简单、快速有效的辐射源识别系统,具有一定的推广价值。Aiming at the present problem of radar radiating source recognition system a new method of radiating resource recognition connecting neural network technology is proposed. All kinds of radar radiating source can be recognized quickly and accurately by this method not only based on the quickness of statistical analyzing radiating source recognition system but also the adaptability and accuracy of its expert system. The practical simulation results show that it is effective for the method, especially its recognition rate and believability with imcomplete, aberrant distorted parameters are raised highly. A radiating source recognition system with simple structure, quickness and effectiveness is designed based on the given method that possesses definete popularized value.

关 键 词：模式识别 神经网络 专家系统 辐射源识别系统 系统设计 雷达辐射源信号 

分 类 号：TN974[电子电信—信号与信息处理] TP391.41[电子电信—信息与通信工程]