宽带侦察信号检测中的门限算法  

Threshold Methods for Signal Detection in Broadband Reconnaissance

作　　者：王永明[1,2] 王世练[1] 张尔扬[1] 程巧金 

机构地区：[1]国防科技大学电子科学与工程学院 [2]解放军94973部队 [3]解放军95920部队

出　　处：《现代雷达》2010年第3期57-61,共5页Modern Radar

摘　　要：利用接收频谱不同噪声子带的独立同分布特性,讨论了通信侦察中信号检测的自适应门限算法。以接收信号DFT输出幅度平方作为统计量时,按指数分布特征计算门限;以DFT幅度平方的积累值作为统计量时,按卡方分布或恒虚警处理计算门限。根据不同统计量服从某一概率分布的特性,上述方法可统一为基于噪声谱线直方图统计的方法,并针对有色噪声频谱提出了一种基于形态学开运算的谱线处理方法。理论分析与仿真实验验证了算法的有效性。Based on independent and identically-distributed noise sub-bands of received spectrum,several methods are discussed for adaptive signal-detection threshold in communication reconnaissance.For DFT squared amplitude,the threshold is computed according to its exponential distribution characteristic,and for the accumulation of DFT squared amplitude,the threshold is computed according to its Chi-square distribution characteristic as well as constant false alarm probability processing.For the probability distribution characteristic of different forms of DFT squared amplitude,methods proposed above can be implemented according to the histogram of noise spectrum lines.In addition,a straightforward way for spectrum lines correction is proposed based on morphological open operation.Analysis and simulation results show the effectiveness of the proposed methods.

关 键 词：通信电子战 信号检测 自适应门限 

分 类 号：TN957.51[电子电信—信号与信息处理]