README

This is the implementation of paper “Deep Learning Based Pilot-free Transmission: Error Correction Coding for Low-resolution Reception under Time-varying Channels”.

0、Citation

@ARTICLE{10179967,
  author={Zeng, Rui and Lu, Zhilin and Zhang, Xudong and Wang, Jintao},
  journal={IEEE Transactions on Vehicular Technology}, 
  title={Deep Learning Based Pilot-free Transmission: Error Correction Coding for Low-resolution Reception under Time-varying Channels}, 
  year={2023},
  volume={},
  number={},
  pages={1-11},
  doi={10.1109/TVT.2023.3294672}}

1、Requirements

This project involves the joint programming of MATLAB and Python, and thus package matlab.engine is need for getdata.py (generate data) and system_main.py (test bit error rate).

Find the path like MATLAB\R2017b\extern\engines\python in your MATLAB installation path, and execute the command python setup.py install.

other requirements:

pytorch >= 1.7.1
torchvision >= 0.8.2
python >= 3.6

2、Tree Arrangement

home
├── data
│   ├── flat/
│   ├── channelA/
│   ├── channelB/
├── models 
│   ├── *Net.py  (all networks) 
│   ├── quantization.py
├── setting 
│   ├── settings.py           (random seed, gpu)
├── tools
│   ├── logger.py
│   ├── parse.py
│   ├── utils.py
├── traditional
│   ├── ofdm.py               (ofdm simulation: LS/MMSE)
│   ├── channel.py
│   ├── match_filtering.py
│   ├── pulse_shaping.py
│   ├── r_filter.py           (Root sign rising cosine function)
├── train.py
├── system_main.py
├── Pilot_*_*                 (Pilot generated by ofdm.py)
...

3、Pipline

A. Train Network

Command:

python train.py -curve TFECCNet_SEAttn -channel channelB -snr 35.0 -qua_bits 1 -lr 1e-4 -modem_num 4

Modifiable Parameters:

• batch_size: batch size
• G: G-fold symbold extension
• N: Block of symbols
• K: preamplification dimensions
• modem_num: modulation order (4 for QPSK, and 16 for 16QAM)
• unit_T: unit increment of soft quantization
• lr: learning rate of Autoencoder
• qua_bits: quantization bits. 0 for unquantization.
• snr: Eb/n0 when training
• curve: select the network type
• channel: select the channel type

B. BER measurement

Command:

python system_main.py -curve TFECCNet_SEAttn -channel channelB -qua_bits 1 -ber_len 50 -modem_num 4 -snr_start 0 -snr_end 35 -snr_step 5

Modifiable Parameters:

• ber_len: ber_len × len is final testing length
• batch_size: batch size
• G: G-fold symbold extension
• N: Block of symbols
• K: preamplification dimensions
• modem_num: modulation order (4 for QPSK, and 16 for 16QAM)
• qua_bits: quantization bits. 0 for unquantization.
• snr_start: start of Eb/n0 when testing
• snr_step: step of Eb/n0 when testing
• snr_end: end of Eb/n0 when testing
• curve: select the network type
• channel: select the channel type

Note:

If $curve=='OFDM_LS/OFDM_MMSE', the qua_bits will not take effect.