Method Of Signal Modulation Recognition Based On Wavelet Transform

Digital communication signal modulation recognition identification has become an important part of modern wireless communication systems. This thesis does research into classifier design method based on the feature parameter extraction.Firstly, this paper makes an analysis of the theories of the wavelet transform. Comparing with the Fourier transform, the characteristics and advantages of the wavelet transform are extracted. Then, wavelet is employed to transform digital modulation signals and meanwhile, the characteristic parameters are extracted as the basis for recognition of signal modulation.Finally, this thesis designs a digital signal modulation recognition method based on neural network, and makes a detail discussion of the network structure and training algorithm of the multilayer perception and meanwhile, the error performance curve is trained by the application of adaptive BP algorithm, as well as signal recognition rate is worked out through lots of experiments on sampling data in different signals to noise ratio and achieve good effect.