摘要
在雷达工作模式识别中,侦察设备对雷达信号参数的测量误差严重影响了识别效果,针对这一问题文中提出时域模糊决策融合(TFDF)的雷达工作模式识别方法。首先分析了雷达脉冲组特征,在脉冲组层次提取雷达信号的脉冲组描述字(PGDW);然后基于联合隶属度函数改进传统的神经网络硬判决方式,实现雷达信号识别的模糊决策;最后运用DS证据理论将多个时刻的模糊决策融合,从而完成雷达工作模式的最终识别。基于时域模糊决策融合的识别方法可以有效改善参数测量误差对工作模式识别效果的影响。仿真结果以及对比实验表明,文中所提算法具有更好的抗噪性能,在雷达信号的参数测量误差为15%时仍具有90%以上的识别率。
Aimed at the problems that in radar operation modes recognition,the errors measured of radar signal parameters of reconnaissance equipment seriously affect recognition effect,a time-domain fuzzy decision fusion(TFDF)recognition algorithm is proposed.Firstly,the characteristics of radar pulse group are analyzed and radar signal pulse group description(PGDW)words at the level of pulse group are extracted.Then,the traditional hard decision of neural network is improved based on the combined membership function,realizing fuzzy decision of radar signal.Finally,the DS evidence theory is used to fuse the fuzzy decision results of multiple moments in time domain,so as to complete the radar operation modes final recognition.The influence of recognition effect because of parameters measure error in radar operation modes recognition is improved based on the method time-domain fuzzy decision fusion.The simulation results and contrast experiment show that the method is good at anti-noise performance,90 percent of the recognition accuracy remains when the parameter errors of measurement are 15%.
作者
董晓璇
程嗣怡
周一鹏
王玉冰
DONG Xiaoxuan;CHENG Siyi;ZHOU Yipeng;WANG Yubing(Aeronautics Engineering College,Air Force Engineering University,Xi 'an710038,China)
出处
《空军工程大学学报（自然科学版）》
CSCD
北大核心
2018年第6期59-65,共7页
Journal of Air Force Engineering University(Natural Science Edition)
基金
航空科学基金(20152096019)
关键词
雷达
神经网络
模糊决策
融合识别
时域
radar
neural network
fuzzy decision
fusion recognition
time-domain
