基于混沌和神经网络的AM调制信号参数检测

数字示波器不能够直接测量AM 调制信号，更不能测量微弱的AM 调制信号。本文运用混沌和Elman神经网络作为检测模型，通过混沌时间序列中重构相空间，获得嵌入维数，确定神经网络结构，并采用单步预测混沌时间序列的方法，实现混沌背景下AM 调制信号参数检测。实验说明，该方法逼近目标精度高，网络适应能力强，从混沌背景中检测出AM 调制信号的周期和幅度，误差小。将上述方法运用到数字示波器设计中，可增加数字示波器测量微弱AM 信号的功能。
关键词：混沌；神经网络；AM 信号；数字示波器；测量；
Abstract: Digital oscilloscope neither measures AM signal, nor weak AM signal. This paper described a method of using Elman neural network to achieve modulation signal parameters detection on chaos background. At first, reconstructed phase space on the time sequence of the chaotic background, and obtained the input dimension by lorenz equation, then confirmed network structure of the neural network, last, forecasted chaotic time series by the single-step forecasting methods. This method has properties of high approaching precision, high ability of network adaptive, high stability, not only can detect transient signal, periodic signal from the chaos background, but also can accurately detect the target signal the shape, including amplitude and period etc. These methods will be applied to the design of oscilloscopes, digital oscilloscope measurements can be realized weak signal functions.
Keywords: Chaos; Elman Networks; AM signal; Digital Oscilloscope; measure;