A lightweight IQ signal modulation recognition method
A lightweight IQ signal modulation recognition method was proposed to address the issues of high computational complexity,excessive model parameters,and slow processing speed in tradi-tional methods for identifying complex signal modulation types.This method constructs one-dimensional networks for the I and Q channels to extract heterogeneous features,and combines dif-ferent sizes of convolution kernels to perceive the multi-scale features of signals,effectively identi-fying different signal modulation types and improving recognition accuracy.Experimental results show that on two public datasets,the overall recognition rates of this method range from 62.44%to 64.61%for signal-to-noise ratios between-20dB and-18dB,and exceed 92%when the signal-to-noise ratio is above 0dB.
NETL
NSTL
万方数据
