5G MIMO Data for Machine Learning: Application to Beam-Selection using Deep Learning

TL;DR

This paper introduces a new methodology combining traffic and raytracing simulators to generate large 5G channel datasets for deep learning-based beam selection in vehicle-to-infrastructure scenarios.

Contribution

It presents a novel data generation approach for 5G MIMO channels, enabling deep learning applications in beam selection with mobility considerations.

Findings

Deep learning models successfully used for beam selection tasks.

Generated datasets improve the training of ML models in 5G scenarios.

Methodology supports classification, regression, and reinforcement learning.

Abstract

The increasing complexity of configuring cellular networks suggests that machine learning (ML) can effectively improve 5G technologies. Deep learning has proven successful in ML tasks such as speech processing and computational vision, with a performance that scales with the amount of available data. The lack of large datasets inhibits the flourish of deep learning applications in wireless communications. This paper presents a methodology that combines a vehicle traffic simulator with a raytracing simulator, to generate channel realizations representing 5G scenarios with mobility of both transceivers and objects. The paper then describes a specific dataset for investigating beams-election techniques on vehicle-to-infrastructure using millimeter waves. Experiments using deep learning in classification, regression and reinforcement learning problems illustrate the use of datasets generated with the proposed methodology