论文部分内容阅读
A method for predicting colored noise by introducing prediction of nonlinear timeseries is presented. By adopting three kinds of neural networks prediction models, the colored noiseprediction is studied through changing the filter bandwidth for stochastic noise and the samplingrate for colored noise. The results show that colored noise can be predicted. The prediction errordecreases with the increasing of the sampling rate or the narrowing of the filter bandwidth. If theparameters are selected properly, the prediction precision can meet the requirement of engineeringimplementation. The results offer a new reference way for increasing the ability for detecting weaksignal in signal processing system.
A method for predicting colored noise by introducing three kinds of neural networks prediction models, the colored noise prediction is studied through changing the filter bandwidth for stochastic noise and the sampling rate for colored noise. The results show that colored The prediction errordecreases with the increasing of the sampling rate or the narrowing of the filter bandwidth. If theparameters are selected properly, the prediction precision can meet the requirement of engineeringmple. The results offer a new reference way for increasing the ability for detecting weaksignal in signal processing system.