dianixn / icc_2023

Code for Achieving Robust Generalization for Wireless Channel Estimation Neural Networks by Designed Training Data

PLEASE DO REALIZE THERE IS A IMPLEMENTATION DELAY OF 7 SAMPLES FOR BOTH MATLAB LTE AND CDL CHANNELS!! PLEASE READ THROUGH THE OFFICIAL MATLAB PAGES.

ICC_2023

Code for "Achieving Robust Generalization for Wireless Channel Estimation Neural Networks by Designed Training Data" accepted by ICC 2023.

Cite as

@INPROCEEDINGS{10279223,
author={Luan, Dianxin and Thompson, John},
booktitle={ICC 2023 - IEEE International Conference on Communications}, 
title={Achieving Robust Generalization for Wireless Channel Estimation Neural Networks by Designed Training Data}, 
year={2023},
volume={},
number={},
pages={3462-3467},
keywords={Training;Wireless communication;Simulation;Design methodology;OFDM;Neural networks;Memory management;Channel estimation;generalization;attention mechanism;deep learning;orthogonal frequency-division multiplexing (OFDM)},
doi={10.1109/ICC45041.2023.10279223}}

The designed training and validation dataset for default pattern are provided in https://datasync.ed.ac.uk/index.php/s/HZ1Xry9TUeDxK2P. Passcode: ICC2023

Please feel free to test and use this training dataset which can achieve robust generalization.