Enhancing URLLC Resiliency in Open RAN Access Networks via AI and Intelligent RIC Architectures

Arnaz, Ava Azadeh

Enhancing URLLC Resiliency in Open RAN Access Networks via AI and Intelligent RIC Architectures

Arnaz, Ava Azadeh

Permalink

Publication Type:: Thesis
Issue Date:: 2025

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download thesisAdobe PDF (4.97 MB)

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Arnaz, Ava Azadeh
dc.date.accessioned	2026-04-10T02:19:06Z
dc.date.available	2026-04-10T02:19:06Z
dc.date.issued	2025
dc.identifier.uri	http://hdl.handle.net/10453/194629
dc.description	University of Technology Sydney. Faculty of Engineering and Information Technology.	en_US.UTF-8
dc.description.abstract	The emergence of Ultra-Reliable Low-Latency Communication (URLLC) applications has introduced unprecedented demands on wireless networks. Open Radio Access Network (Open RAN) architecture offers promising capabilities through its Near Real-Time RAN Intelligent Controller (Near-RT RIC), yet critical challenges remain in supporting URLLC reliability requirements. This thesis presents innovative solutions for enhancing Near-RT RIC for URLLC applications. First, we develop a multi-objective optimization framework that enhances RAN control functions with Pareto-optimal decision making. Traditional approaches prioritize single objectives and handle failures through costly retries. We created HORLA (HandOver Reinforcement Learning Application) to simultaneously optimize multiple parameters for both reliability and performance. Experimental validation demonstrates a 40% reduction in handover failures compared to traditional approaches, while maintaining sub-second latency and reducing energy consumption by 57%. Results confirm that multi-objective controllers are essential for achieving necessary reliability within near-real-time constraints. Second, we present a security study addressing vulnerabilities in AI-enabled Near-RT RIC systems through investigations of reward manipulation, last-layer distortion, and parameter tampering. Our work demonstrates how sophisticated attacks can compromise network performance while evading traditional monitoring systems. Third, we propose PULSE (Predictive Ultra-reliable Low-latency System Engine), a Near-RT RIC xApp that redefines reliability solutions by incorporating semantic-aware processing. By extending Shannon’s communication theory, PULSE leverages transformer-based understanding to reconstruct lost packets without retransmissions. Our implementation achieves 100% prediction accuracy for up to 10% packet loss and 93.96% accuracy with 10-50% loss, while delivering submillisecond processing times. Finally, we introduce DANTE (Drone Adaptive Natural-to-Encoded Text Engine), a Near-RT RIC xApp that moves command standardization from endpoints to the centralized RAN edge. DANTE achieves 98.90% accuracy transforming natural language commands into standardized formats while maintaining strict latency requirements, improving efficiency and reliability across multiple devices. These contributions establish a new paradigm for near-real-time applications’ reliability in Open RAN. Our thesis demonstrates the necessity for innovative frameworks in Near-RT RIC that enhance reliability while maintaining strict latency requirements, providing a foundation for future research in intelligent wireless networks for next-generation URLLC applications.	en_US.UTF-8
dc.format	Thesis (PhD)
dc.language.iso	en_US	en_US.UTF-8
dc.rights	info:eu-repo/semantics/openAccess
dc.rights	The author owns the copyright in this thesis including all reproduction and reuse rights for the work. The work may not be altered without the permission of the copyright owner. Attribution is essential when quoting or paraphrasing from this thesis.
dc.rights	© 2025 Ava Azadeh Arnaz
dc.rights	au.edu.uts.lib/nph
dc.title	Enhancing URLLC Resiliency in Open RAN Access Networks via AI and Intelligent RIC Architectures	en_US.UTF-8
dc.type	Thesis
utslib.copyright.status	open_access	*

Abstract:

The emergence of Ultra-Reliable Low-Latency Communication (URLLC) applications has introduced unprecedented demands on wireless networks. Open Radio Access Network (Open RAN) architecture offers promising capabilities through its Near Real-Time RAN Intelligent Controller (Near-RT RIC), yet critical challenges remain in supporting URLLC reliability requirements. This thesis presents innovative solutions for enhancing Near-RT RIC for URLLC applications. First, we develop a multi-objective optimization framework that enhances RAN control functions with Pareto-optimal decision making. Traditional approaches prioritize single objectives and handle failures through costly retries. We created HORLA (HandOver Reinforcement Learning Application) to simultaneously optimize multiple parameters for both reliability and performance. Experimental validation demonstrates a 40% reduction in handover failures compared to traditional approaches, while maintaining sub-second latency and reducing energy consumption by 57%. Results confirm that multi-objective controllers are essential for achieving necessary reliability within near-real-time constraints. Second, we present a security study addressing vulnerabilities in AI-enabled Near-RT RIC systems through investigations of reward manipulation, last-layer distortion, and parameter tampering. Our work demonstrates how sophisticated attacks can compromise network performance while evading traditional monitoring systems. Third, we propose PULSE (Predictive Ultra-reliable Low-latency System Engine), a Near-RT RIC xApp that redefines reliability solutions by incorporating semantic-aware processing. By extending Shannon’s communication theory, PULSE leverages transformer-based understanding to reconstruct lost packets without retransmissions. Our implementation achieves 100% prediction accuracy for up to 10% packet loss and 93.96% accuracy with 10-50% loss, while delivering submillisecond processing times. Finally, we introduce DANTE (Drone Adaptive Natural-to-Encoded Text Engine), a Near-RT RIC xApp that moves command standardization from endpoints to the centralized RAN edge. DANTE achieves 98.90% accuracy transforming natural language commands into standardized formats while maintaining strict latency requirements, improving efficiency and reliability across multiple devices. These contributions establish a new paradigm for near-real-time applications’ reliability in Open RAN. Our thesis demonstrates the necessity for innovative frameworks in Near-RT RIC that enhance reliability while maintaining strict latency requirements, providing a foundation for future research in intelligent wireless networks for next-generation URLLC applications.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/194629