[Paper Review] Intelligent O-RAN for Beyond 5G and 6G Wireless Networks

Solmaz Niknam, Abhishek Roy+6 more|arXiv (Cornell University)|May 17, 2020

Advanced Wireless Communication Technologies68 citations

TL;DR

The paper demonstrates an intelligent congestion management scheme for O-RAN that uses LSTM-based traffic prediction to trigger cell splitting, evaluated on real Mumbai network data.

ABSTRACT

Building on the principles of openness and intelligence, there has been a concerted global effort from the operators towards enhancing the radio access network (RAN) architecture. The objective is to build an operator-defined RAN architecture (and associated interfaces) on open hardware that provides intelligent radio control for beyond fifth generation (5G) as well as future sixth generation (6G) wireless networks. Specifically, the open-radio access network (O-RAN) alliance has been formed by merging xRAN forum and C-RAN alliance to formally define the requirements that would help achieve this objective. Owing to the importance of O-RAN in the current wireless landscape, this article provides an introduction to the concepts, principles, and requirements of the Open RAN as specified by the O-RAN alliance. In order to illustrate the role of intelligence in O-RAN, we propose an intelligent radio resource management scheme to handle traffic congestion and demonstrate its efficacy on a real-world dataset obtained from a large operator. A high-level architecture of this deployment scenario that is compliant with the O-RAN requirements is also discussed. The article concludes with key technical challenges and open problems for future research and development.

Motivation & Objective

• Introduce Open RAN (O-RAN) concepts and the role of openness and intelligence in next-generation networks.
• Propose an ML-enabled radio resource management scheme for congestion prediction and mitigation within O-RAN.
• Validate the approach on a real-world Mumbai LTE dataset and map the solution to O-RAN control loops and interfaces.
• Discuss deployment architectures, implementation steps, and practical challenges for intelligent RAN.

Proposed method

• Employ an LSTM recurrent neural network to learn and predict temporal traffic patterns and potential congestion.
• Define congestion as: average user-perceived IP throughput < 1 Mbps AND DL-PRB utilization > 80%.
• Train the LSTM model on a real-world Mumbai cellular dataset (17 eNBs, 18 cells each, 25 days) using 2 layers of 12 LSTM units.
• Deploy the trained model at non-RT RIC for training and near-RT RIC for inference within the O-RAN framework.
• Trigger congestion mitigation via cell splitting and, where applicable, dual connectivity, based on predicted congestion.
• Describe end-to-end deployment flow aligning ML training/inference with O1/A1/E2 interfaces and xAPP CPM.

Experimental results

Research questions

• RQ1 Can LSTM-based traffic prediction accurately forecast congestion in a dense urban RAN?
• RQ2 Does preemptive cell splitting based on ML predictions improve user-perceived throughput and DL-PRB utilization?
• RQ3 How can an ML-enabled congestion management loop be integrated with O-RAN interfaces (O1, A1, E2) and xAPPs?
• RQ4 What deployment architectures and data workflows support intelligent RAN in beyond-5G/6G scenarios?

Key findings

• Prediction accuracy of the LSTM model averaged 92.64% for traffic parameters.
• Preemptive cell splitting based on congestion predictions significantly improves user-perceived IP throughput.
• DL-PRB utilization and throughput metrics respond to ML-driven congestion relief within the defined KPI targets.
• The proposed approach is mapped to O-RAN control loops with ML training at non-RT RIC and inference at near-RT RIC.
• A high-level deployment architecture shows data collection, ML training/inference, CPM xAPP, and E2-based action deployment.