混沌背景下基于小波神经网络的弱信号检测

依据Takens嵌入定理提出了一种基于小波神经网络（WNN）的强混沌背景中微弱信号的检测方法。该方法利用混沌系统的单变量值对混沌背景重构相空间，采用小波神经网络所具有的强大的学习能力和非线性处理能力建立了混沌背景噪声的一步预测模型，使其与混沌背景噪声具有相同的基本动力学特征，并通过设定合适的预测误差门限来检测掩埋在混沌背景中的有用微弱信号。仿真结果验证了该方法的可行性。

Detecting Weak Signal from Chaotic Background with Wavelet Neural Network

Weak signal detection is a important study aspect in target detection.This paper presents a wavelet neural network method(WNN)to detect weak harmonic signal embedded in chaotic background.Based on embedding theory,the method utilizes the observed values of single variable of chaotic system to reconstruct phase space.Based on the wavelet neural network powerful ability of studying and nonlinear processing build one-step predictive model which has the same underlying dynamics as the chaotic background.then set an appropriate threshold of the prediction errors to detect weak signal buried in strong chaotic noise.The results show the feasibility of the method.