[cosc-grad-students-list] Abdullah Anjum Thesis Defense Presentation

[Eulenfeld, Menda]

Student Name: Abdullah Anjum
Program: Master of Science, Computer Science
Committee Chair: Dr. Mehdi Sookhak

Thesis Title:
NT&TP: Network Traffic and Trajectory Prediction in V2X-Enabled O-RAN Systems Through Transformer Learning

Abstract:
Vehicle-to-Everything (V2X) communication systems are facing critical challenges to ensure the Quality of Service (QoS) of dynamic vehicles moving through urban environments across distributed Radio Service Units (RSUs). Current reactive approaches only respond to network changes after they have occurred, resulting in frequent handoff failures and service degradation. Proactive management of resources requires the ability to predict vehicle trajectories with accuracy to anticipate movements and plan the network accordingly. This thesis proposes a Transformer-based Network Traffic and Trajectory Prediction (NT&TP) model in O-RAN to enable accurate multi-horizon forecasting of vehicle movements and network dynamics. Using trajectories of 1,386 vehicles with 9.87 million records from a realistic urban traffic simulation in the city of Corpus Christi, Texas, we show that our approach outperforms traditional LSTM/GRU baselines for multiple prediction horizons. The predicted trajectories support applications such as proactive resource allocation in the O-RAN RAN Intelligent Controller (RIC). This thesis delivers a generalizable framework, a large-scale V2X trajectory dataset, and open-source implementations for the advancement of intelligent transportation systems and next-generation wireless networks.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.tamucc.edu/pipermail/cosc-grad-students-list/attachments/20260414/87fde046/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/calendar
Size: 6828 bytes
Desc: not available
URL: <http://listserv.tamucc.edu/pipermail/cosc-grad-students-list/attachments/20260414/87fde046/attachment.ics>