Machine Learning Modeling of Reliable Railway Communication: Generating Packet Replication Mode Delays

Loading...
Thumbnail Image

URL

Journal Title

Journal ISSN

Volume Title

School of Electrical Engineering | Master's thesis

Department

Mcode

Language

en

Pages

112

Series

Abstract

Finland is upgrading its railway communication system by utilizing public 4G and 5G mobile networks through the Digirail project. To assess the performance of these networks, a comprehensive measurement campaign was conducted across Finland’s railway network. Initially, measurements were performed in Best Quality Mode across the entire network. However, since this mode introduces artificial delays and does not accurately reflect real network conditions, Packet Replication Mode measurements were subsequently carried out on selected sections to provide a more realistic assessment of network performance. Given that Packet Replication Mode data is limited to specific sections, this thesis aims to model network delays based on the measurements obtained in Packet Replication Mode and to use this model to generate a comprehensive dataset estimating delays across Finland’s entire railway network. The resulting dataset offers a more accurate representation of network performance, thereby supporting future developments in railway communication systems and contributing to the safe and efficient operation of railway control.

Description

Supervisor

Mähönen, Petri

Thesis advisor

Sigg, Stephan
Shamekh, Mohamed

Other note

Citation