摘要
根据海杂波具有混沌特性这一先验信息,利用RBF神经网络预测器的学习和对非线性函数的逼近能力,首先重构海杂波的内在动力学,然后引入一种基于预测误差的检测方法对微弱脉冲信号和矩形信号进行检测分析,最后得出了这种检测方法对微弱脉冲信号有较好的检测性能,信杂比改善可达43.3898dB;而对矩形信号只在信号的起点和终点附近提高了检测性能,有类似的信杂比改善,中间各点检测不到信号,信杂比改善不大。
Based on the fact that sea clutter is chaotic which is known to us, the ability of study and prediction of the RBF neural network is utilized to reconstruct the underlying dynamics of sea clutter data firstly in this paper. Then detection analysis using weak pulse signal and rectangle signal is carried out by using of a prediction error-based method. This method can detect weak pulse signal with SCR improvement to 43.389 8 dB. However, detection performance improves only nearby the beginning, the end and the similar SCR improvement can be gotten at these two points, points between them can not be detected and can not be gotten similar SCR improvement.
出处
《微型机与应用》
2010年第5期45-48,共4页
Microcomputer & Its Applications
关键词
海杂波
混沌
神经网络
目标检测
sea clutter
chaos
RBF neural network
target detection
