【摘要】 基于改进的经验模式分解,提出了多分量LFM雷达辐射源信号的分析方法.该方法用RBF神经网络对端点延拓削弱边界效应,将自相关函数与相关系数结合估计分量的数量,通过模式分解滤波和平均滑动消除噪声影响,以提高算法的分解精度.理论分析和实验表明,在较宽的信噪比范围内,使用该方法能够正确提取各分量信号的瞬时频率和有效地估计多分量LFM辐射源信号的分量数量.更多还原

【Abstract】 A method for detection of multi-component LFM radar emitter signals was proposed based on empirical model decomposition(EMD).In this method,to improve decomposition precision,a forecast method of radial basis function(RBF) network is used to suppress the end effect of EMD,the correlation coefficient and auto-correlation function are used to estimate the component number of radar signals,and an EMD filter and mean value glide are used to eliminate the effects of noises.Theoretical analysis and experimental results show that this method can accurately Abstract instantaneous frequency of every component of LFM radar signals and estimate their component number.更多还原