Sub-6 Predicts mmWave Beam-forming Vectors:

This is an implementation of the Deep Learning (DL) solution that uses sub-6 GHz channels to predict top-n beams of mmWave users. With the approperiate modifications and given the right dataset, it could also be used to generate all the figures in Deep Learning for mmWave Beam and Blockage Prediction Using Sub-6GHz Channels.

Requirements:

Essential:

1- MATLAB deep learning toolbox.

2- DeepMIMO dataset

Optional:

1- NVIDIA GPU card.

2- CUDA toolkit.

3- cuDNN package.

Running Instructions:

1- Generate the datasets using scenarios O1_28 and O1_3p5 in the DeepMIMO dataset. Use the parameters listed in Table.1, Section VII-B of the paper.

2- Prepare two MATLAB structures, one for sub-6GHz data and the other for 28GHz. Please refer to the comments at the beginning of main.m for more information on the data structures.

3- Assign the paths to the two MATLAB structures to the two parameters: options.dataFile1 and options.dataFile2 in the beginning of main.m.

4- Run main.m to get the figure 4-b in the paper.

REMARK: Transmit power range is defined in tx_power in main.m.

Citation:

If you use these codes or a modified version of them, please cite the following work:

@ARTICLE{Alrabeiah2020,
  author={Alrabeiah, Muhammad and Alkhateeb, Ahmed},
  journal={IEEE Transactions on Communications}, 
  title={Deep Learning for mmWave Beam and Blockage Prediction Using Sub-6 GHz Channels}, 
  year={2020},
  volume={68},
  number={9},
  pages={5504-5518},
  doi={10.1109/TCOMM.2020.3003670}}

License:

This code package is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License