Main_AMC_3_1.m -> Main file
The AMC_EWT project contains 3 stages
-
Signal Generation
We have considered 9 modulation types for our Classifiaction out of which 6 are digital and 3 are analog modulation types. The random signals are generated and passed through the Modulators which are defined as modulation functions and then passed through various Noise channels.
audio_mix_441.wav -> The Audio file used for analog signals.
getSource.m -> The random Signal Genertion.
getModulator.m -> The function for picking up specified modulator.
helperModClassFrameStore.m -> Creates a frame store object to store the modulated signals.
helperModClassFrameGenerator -> Removes transients from the signal, trims to spefied size and nirmalize the signal to Generate frames for machine learning.
The Modulation functions are defined in following functions:
- bfmModulator.m
- bpskModulator.m
- cpfskModulator.m
- dsbamModulator.m
- gfskModulator.m
- pam4Modulator.m
- psk8Modulator.m
- qam16Modulator.m
- qam64Modulator.m
- qpskModulator.m
- ssbamModulator.m
Please Refer - https://in.mathworks.com/help/deeplearning/ug/modulation-classification-with-deep-learning.html
- Signal Decomposition using FBR-EWT:
EWT_Meyer_FilterBank.m -> Created a filter bank based on the boundaries(set of frequency segments).
EWT_Meyer_Scaling.m -> Generate the 1D Meyer wavelet in the Fourier domain associated to the segment.
Please Refer - https://in.mathworks.com/matlabcentral/fileexchange/42141-empirical-wavelet-transforms
- Classification using Deep CNN:
Please Refer - https://in.mathworks.com/help/deeplearning/ug/create-simple-deep-learning-network-for-classification.html