Agentic Systems for O-RAN

An agentic system is a technological framework often powered by artificial intelligence in which an "agent" (such as an AI model or software entity) operates autonomously to pursue complex, multi-step goals on a user's behalf. Unlike standard chatbots that merely generate text, agentic systems can perceive their environment, reason through tasks, make decisions, and execute actions by using external tools and APIs to achieve a specific objective .

Open Radio Access Network (O-RAN) is transforming mobile network architecture by introducing disaggregated, virtualized, and cloud-native network functions. The adoption of artificial intelligence, particularly Large Language Models (LLMs) and agentic AI systems, presents new opportunities to automate network management, improve operational efficiency, and enable intelligent decision-making across O-RAN deployments. These technologies have the potential to support network monitoring, anomaly detection, fault diagnosis, resource optimization, and closed-loop automation in increasingly complex communication networks.

Despite these advancements, building reliable AI-driven systems for O-RAN remains a significant research challenge. Agentic systems must reason over heterogeneous telemetry data, understand dynamic network environments, coordinate across multiple network components, and provide trustworthy recommendations while satisfying the performance and reliability requirements of communication networks. Developing scalable, explainable, and robust AI solutions that can effectively assist network operators is an active area of research for future 5G and 6G networks.

Research Problems

Problem	Description
Agentic AI for Network Automation	Developing intelligent agents capable of reasoning over network telemetry and autonomously supporting network management and operational workflows.
Anomaly Detection & Root Cause Analysis	Designing AI-based methods to detect anomalies, identify their underlying causes, and provide actionable insights for communication networks.
AI-assisted Network Optimization	Applying machine learning and agentic systems to optimize network performance, resource utilization, and service quality in dynamic O-RAN environments.
Trustworthy & Explainable AI	Building transparent, reliable, and explainable AI systems that improve operator trust and enable safe deployment in production networks.

Research Scholars

Yaswanth Kumar L. S., Michael Suguna Kumar Victor, Harshwardhan Gaikwad