"CHANNEL SELECTION OF A HYBRID COMMUNICATION SYSTEM USING MACHINE LEARN" by Julie C. Smith
 

Electrical and Computer Engineering ETDs

Publication Date

Fall 10-28-2024

Abstract

As the need for “data on demand” increases, users are moving to hybrid communication systems to increase data throughput. Users in both commercial and government sectors are looking to integrate laser communication (lasercom) and W/V band (81-86 GHz / 71-76 GHz) solutions onto their platforms. These systems are subject to attenuation due to weather and atmospheric conditions. Employing a hybrid communication system requires knowledge of which system to utilize more heavily in given atmospheric conditions. This research investigates means to develop channel prediction algorithms for hybrid systems using weather imagery. Convolutional Neural Networks (CNNs) were adapted and trained with these weather products and then utilized to develop a decision algorithm that can predict which system will provide the highest performance in a given atmospheric condition. The decision algorithms were developed and tested on a terrestrial link provided by the Air Force Research Laboratory, Space Vehicles Directorate (AFRL/RV).

Keywords

Laser communications, convolutional neural nets, hybrid communication systems

Document Type

Dissertation

Language

English

Level of Degree

Doctoral

Department Name

Electrical and Computer Engineering

First Committee Member (Chair)

Christos Christodoulou

Second Committee Member

Jane Lehr

Third Committee Member

Trilce Estrada

Fourth Committee Member

Richard Scott Erwin

Fifth Committee Member

Thomas Farrell

Download

Share

COinS