tristantengrp / itu_ml5g_phy_ncsu_mmwavenet

ITU Artificial Intelligence/Machine Learning in 5G Challenge Site-Specific Channel Estimation with Hybrid MIMO Architectures Neel Kanth Kundu, Nilesh Kumar Jha, and Amartansh Dubey Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong Email: {nkkundu, nkjha, adubey}

This repository contains the solution to ITU Artificial Intelligence/Machine Learning in 5G Challenge

Problem Statement: Site-Specific Channel Estimation with Hybrid MIMO Architectures (https://research.ece.ncsu.edu/ai5gchallenge/)

Team Name: mmWaveNet

Organization: The Hong Kong University of Science and Technology (HKUST)

Members:

1. Neel Kanth Kundu ([email protected])
2. Nilesh Kumar Jha ([email protected])
3. Amartansh Dubey ([email protected])

Report

report

Test Results

Test Results

How to Use:

1. Download the Dataset
2. Download gen_channel_ray_tracing_rev.m from https://research.ece.ncsu.edu/ai5gchallenge/#datasets
3. For Training: Use line search to find L for training SNR -15dB, -10dB and 0dB by running Training.m

For Test:

1. First download the test data received pilots and precoder/combiner .mat files from https://research.ece.ncsu.edu/ai5gchallenge/#datasets
2. Run test.mat by changing the variables Dataset_pilots = 20/40/80 and Dataset_snr = 1/2/3 to get the 9 files containing the estimated channels