Regenerative RIC-Based xApp for O-RAN NTN

Abstract

The integration of Non-Terrestrial Networks (NTNs) into beyond-5G and 6G systems enables ubiquitous connectivity and extended coverage, but efficient Radio Resource Management (RRM) remains challenging due to varying channel conditions, long propagation delays, and non-contiguous Physical Resource Block (PRB) allocation. This paper analyzes an xApp-based scheduling framework operating within the near-real-time Radio Access Network (RAN) Intelligent Controller (near-RT RIC) of an Open-RAN (O-RAN) enabled regenerative NTN system. The framework combines SNR-aware prioritization, fairness control, and PRB allocation using Fenwick Tree and sliding-window methods, while a pre-trained Reinforcement Learning (RL) agent performs lightweight inference for adaptive decision-making. Evaluation results demonstrate improvements in throughput, latency, fairness, and resource utilization, while maintaining low computational energy overhead compared to conventional methods.