基于时频回归特性分析的信号脉内调制识别  

Intrapulse Modulation Recognition of Signals Based on Linear Rregression of Time-Frequency Curve

作　　者：华永平[1] 胡国兵[1,2] 刘渝[2] 

机构地区：[1]南京信息职业技术学院电子信息学院,江苏南京210046 [2]南京航空航天大学信息科学与技术学院,江苏南京210016

出　　处：《计算机仿真》2010年第12期12-15,59,共5页Computer Simulation

摘　　要：研究雷达识别优化问题,在电子侦察领域,对所截获信号的脉内调制是介于信号检测与解调之间的重要识别方式。为了提高雷达功能,通过短时傅立叶变换,获取信号的时频曲线,在分析线性回归特性的基础上,将脉内调制识别转化为时频曲线线性回归模型。进行计算机仿真,结果表明,方法在无接收信号的任何先验知识,信噪比4dB时,平均识别正确率达98%,仿真也证明了对常用雷达脉内调制方式可提高识别效率。In the field of radar electronic surveillance,intrapulse modulation recognition of the intercepted signals is an important intermediate operation between signal detection and demodulation.A simple decision-theoretic approach,based on the properties of the linear regression of time-frequency curve（TFC） of the intercepted signal is proposed to recognize the common intrapulse modulation of radar signals.The main advantages of the method are that we simplify the recognition problem to lack of fit test for the linear regression of the TFC.Computer simulation results show that the performance of the proposed method achieves average correct recognition rate of 98% at the SNR of 4dB and it is found to be effective in classifying the common intrapulse modulation radar signals.

关 键 词：脉内调制识别 时频曲线 线性回归 失拟检验 

分 类 号：TN957[电子电信—信号与信息处理]