modulation classification using deep learning

Hi,

I need your help . I am working on Modulation Classification and I found a matlab tutorial about it using CNN classification algorithm.

https://www.mathworks.com/help/deeplearning/ug/modulation-classification-with-deep-learning.html

First I want to generate random data and waveform for each modulation type. I run the trainFirst.m and waveform.m , respectively. I have also getModulator.m, getSource.m and helperModClassFrameGenerator.m

I can generate frames for digital 8 different modulation types. How can I see these data ? In example, It is added to frameStore. I didn't get it ?

Have you any other code about modulation classification ? Bcs this example is complicated according to me. I couldn't follow these steps in the example .

Also I have another question about this example. it uses analyzeNetwork() function and tells that "this network has 98,323 learnables" How do we get this info ?

Answer 1

[Shashank Gupta]

Hi Fatma,

we generally put the data in some datastore so as to have a better control over the processing of data. you can consider frameStore as one such, in the example even signalDatastore is used to control the real and imaginary part of data. You can visaulise them using helperModClassPlotTimeDomain or helperModClassPlotSpectrogram functions. Do explore these datastores, go to their properties and see what all you can do with it, you can also read a batch of data, transform them or even process them easily.

Regarding analyzeNetwork function, it is an helper function to visualise and analyze the deep learning model, if it shows, you have 98,323 learnables means the deep learning model has 98,323 learnable parameter which include weights and biases. the function calculate the learnable in each layer, add them up and show us the total number.

I hope I gave you some insight for you to explore.

Cheers,