End-to-End O-RAN Control-Loop For Radio Resource Allocation in SDR-Based 5G Network
Asheesh Tripathi, Jaswanth S. R. Mallu, Md. Habibur Rahman, Abida Sultana, A. Sathish, Alexandre Huff, Mayukh Roy Chowdhury, Aloizio Pereira Da Silva
Abstract
End-to-end Open Radio Access Network (ORAN) demos, including Software-defined Radios (SDRs), can identify key gaps in the early stages of research and accelerate time-to-market and early adoption for next-generation wireless technologies. This work showcases an innovative end-to-end 5G-based ORAN deployment that leverages open-source tools and an Artificial Intelligence (AI)/Machine Learning (ML) framework. The deployment utilizes Software Radio Systems RAN (srsRAN) Central Unit (CU)-Distributed Unit (DU) connected to the Open5GS core network operating in 5G standalone (SA) mode. It also demonstrates the integration of Non-Real Time RIC (Non-RT RIC) and Near-Real Time RIC (Near-RT RIC) with embedded intelligence to perform radio resource allocation. An AI/ML framework deploys an optimized ML model as an rApp that complements a resource allocation xApp.