The rapid emerging wireless technology changed the way of using digital media completely.
For example, using the smartphone to access online services determines a steady,
reliable and robust network connection. Requirements which push network development,
improvement and further expansion of specifications. But also new problem statements start
to arise as the used technology-type is a shared medium.
All devices are forced to use portions of the available bandwidth and when
sharing a medium, inconsiderate and excessive use thereof should be avoided. Thinking
about mobile networks, measuring the available throughput for one user could be seen
as such an unnecessary usage, and therefore should be avoided or minimized to relieve
the network. Considering the major improvements in the Artificial Intelligence domain
in the last years, the question presents itself: Is it possible to predict the mobile networks
throughput by using a specially trained Neural Network?
The thesis project ”Examining the Feasibility of Using a Neural Network to Predict
Mobile Networks Throughput” provides and reports insights gained by several
experimental attempts to build, train and evaluate a deep learning model which predicts
the available mobile throughput using client-side parameters as input values. These parameters are easily accessible with nearly every mobile device and are used to measure, among
other things, the quality of the connection between a mobile device and a network cell.
By representing the consecutive steps which were taken to achieve the given task, the
thesis provides all relevant information to retrace the examined procedures and decisions.
Providing the basic background knowledge and how the data was prepared for further
use, should allow the reader to follow the thesis to it’s full extent. As the project deals
heavily with Artificial Intelligence, a short overview of the domain should help understand
the used processes and models.
Facts:
Projekt-Team: Sebastian Rieger, Andreas Roth, Hannes Slowak
Projekt-Betreuer: DI(FH) DI Peter Dorfinger
Typ: Bachelorprojekt
Studiengang: Bachelor Informationstechnik & System-Management
