Welcome to the upgraded BRAC University Institutional Repository. We are currently organizing collections after a recent system upgrade. Homepage category counters may temporarily show lower numbers while syncing, but over 27,000 repository items remain safe and accessible. Please use the search bar to find theses, scholarly outputs, and institutional documents.

A comparative study of listen before talk categories using machine learning

Loading...
Thumbnail Image

Publisher

Brac University

Citation

Abstract

The cellular industry is seeking solutions to efficiently utilize the available spectrum band due to the rapid growth of wireless traffic and technological advancements. One potential solution that has gained attention is implementing Long Term Evolution (LTE) with unlicensed spectrum (LTE-U) using a Listen before Talk (LBT) approach, as prescribed by international regulators. To ensure fair channel access for co-located networks, it is crucial to establish a coexistence strategy that incorporates expected traffic requirements for both present and future needs. Machine learning has been recognized for its ability to automate critical wireless communication network activities, gather data from multiple sources, and employ various algorithms. Thus, this project emphasizes the significance of researching LTE and Wi-Fi coexistence in unlicensed spectrum using machine learning. It provides an overview of existing LTE-U and Wi-Fi technologies, and reviews the studies that have been conducted on their coexistence. The project also discusses LBT mechanisms and their categories as defined by the 3GPP standard, as well as previous research conducted in various categories, providing a basis for future research. The study evaluates the performance of each priority class of LBT Cat 4 using the Jains Fairness with machine learning approach to determine the best coexistence priority class of LBT Cat 4 that will enhance future network performance when coexisting with Wi-Fi. Thus in wireless communication systems, machine learning can be used to optimize the LBT protocol by learning the patterns and characteristics of the communication channel. By training the large amounts of data collected from the communication channel, the network can learn to predict when the channel will be free and when it will be busy. This can help to reduce the waiting time for devices and increase the efficiency of the communication system. Moreover it help us to understand the channel sharing fairness and signal detection probability better for each of the priority class of LBT Cat4.

LC Subject Headings

Description

Cataloged from PDF version of the project report.
Includes bibliographical references (pages 60-82).
This project report is submitted in partial fulfilment of the requirements for the degree of Master of Science in Electrical and Electronic Engineering, 2023.

Publisher Link

Type

Project Report