yujie2016 / deepcom

a CNN based end to end communication systems

A CNN based end to end communication systems

Updated: 07/02/2019.
This repository contains source code necessary to reproduce the results presented in the following paper:
A CNN-Based End-to-End Learning Framework Towards Intelligent Communication Systems
by Nan Wu, Xudong Wang, Bin Lin, and Kaiyao Zhang, accepted to IEEE access.

Dependency

• Python (3.7.0)
  
• Numpy (1.15.4)
  
• Keras (2.2.4)
  
• Tensorflow (1.13.1)
  

AWGN channel

• use model_LBC_AWGN.py to train model at a fixed Eb/N0
• use test_model_LBC_AWGN.py to test the model at a range of Eb/N0

Rayleigh fading channel

• use model_LBC_Rayleigh.py to train model at a fixed Eb/N0
• use test_model_LBC_Rayleigh.py to test the model at a range of Eb/N0

Bursty AWGN channel

• use model_LBC_Bursty_AWGN.py to train model at a fixed Eb/N0
• use test_model_LBC_Bursty_AWGN.py to test the model at a range of Eb/N0

Differential Version

• use model_DLBC_Rayleigh.py to train model at a fixed Eb/N0
• use test_model_DLBC_Rayleigh.py to test the model at a range of Eb/N0
  The differential version currently only supports n=1, adding n involves complex multiplication in high-dimensional space，and is under construction

Questions?

if you have any questions, please e-mail([email protected]).