Demodulation of AM signal Based on Adaptive Linear Networks and Elman Networks
This paper proposed a method of combining an adaptive linear network with a spatio-temporal neural network to enhance a feature of digital oscilloscope to measure AM modulation signaI A basic Elman neural network structure was improved by adding the connection of weights from output layer to hidden layer and delayed the feedback signal and adopting learning algorithm of grads descendent with an additive momentum fact.The improved network has fast Iearning speed.high precision of approaching function,and good adaptability,also there is not oscillation or wavelet on the output wavefornl.At meantime,an adaptive linear network and an Eiman network were combined in series.An adaptive linear network realized optimal filtering function and enhanced anti-jamming capability.It realized to demodulate AM signal by combining two networks.This novel method decreases greatly computational burden and the error is very small.It is advantaged to program a function and interpolate into the function database of digital oscilloscopes to realize feature of measurement of AM modulation signal.
Adaptive Linear Networks Elman Networks Digital Oscilloscope Demodulation
Yuan Jimin Li Xiaoling Zhang Xiangyu Li Jianping4
国际会议
The International Conference Information Computing and Automation(2007国际信息计算与自动化会议)
成都
英文
825-829
2007-12-19（万方平台首次上网日期，不代表论文的发表时间）