Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Interference-Aware PMI selection for MIMO systems in an O-RAN scenario

Rawlings Ntassah, Gian Michele Dell'Aera, Fabrizio Granelli

TL;DR

This work tackles PMI selection in dense 5G MU-MIMO networks under inter-cell interference by introducing Inter-A2C, an interference-aware Advantage Actor-Critic xApp designed for O-RAN deployment. The method uses a multi-objective RL formulation with UE prioritization to adjust PMI selections, incorporating both spectral efficiency and interference costs through a tailored state, action, and reward structure. Evaluations in an O-RAN TIM-based simulator show that Inter-A2C yields significant SE gains, lower interference, and more balanced PRB utilization compared with baseline Follow PMI and standard A2C approaches, demonstrating practical viability for real-time PMI management. Overall, the approach highlights the potential of integrated AI/ML control within O-RAN to enhance coordination, efficiency, and user experience in next-generation networks.

Abstract

The optimization of Precoding Matrix Indicators (PMIs) is crucial for enhancing the performance of 5G networks, particularly in dense deployments where inter-cell interference is a significant challenge. Some approaches have leveraged AI/ML techniques for beamforming and beam selection, however, these methods often overlook the multi-objective nature of PMI selection, which requires balancing spectral efficiency (SE) and interference reduction. This paper proposes an interference-aware PMI selection method using an Advantage Actor-Critic (A2C) reinforcement learning model, designed for deployment within an O-RAN framework as an xApp. The proposed model prioritizes user equipment (UE) based on a novel strategy and adjusts PMI values accordingly, with interference management and efficient resource utilization. Experimental results in an O-RAN environment demonstrate the approach's effectiveness in improving network performance metrics, including SE and interference mitigation.

Interference-Aware PMI selection for MIMO systems in an O-RAN scenario

TL;DR

This work tackles PMI selection in dense 5G MU-MIMO networks under inter-cell interference by introducing Inter-A2C, an interference-aware Advantage Actor-Critic xApp designed for O-RAN deployment. The method uses a multi-objective RL formulation with UE prioritization to adjust PMI selections, incorporating both spectral efficiency and interference costs through a tailored state, action, and reward structure. Evaluations in an O-RAN TIM-based simulator show that Inter-A2C yields significant SE gains, lower interference, and more balanced PRB utilization compared with baseline Follow PMI and standard A2C approaches, demonstrating practical viability for real-time PMI management. Overall, the approach highlights the potential of integrated AI/ML control within O-RAN to enhance coordination, efficiency, and user experience in next-generation networks.

Abstract

The optimization of Precoding Matrix Indicators (PMIs) is crucial for enhancing the performance of 5G networks, particularly in dense deployments where inter-cell interference is a significant challenge. Some approaches have leveraged AI/ML techniques for beamforming and beam selection, however, these methods often overlook the multi-objective nature of PMI selection, which requires balancing spectral efficiency (SE) and interference reduction. This paper proposes an interference-aware PMI selection method using an Advantage Actor-Critic (A2C) reinforcement learning model, designed for deployment within an O-RAN framework as an xApp. The proposed model prioritizes user equipment (UE) based on a novel strategy and adjusts PMI values accordingly, with interference management and efficient resource utilization. Experimental results in an O-RAN environment demonstrate the approach's effectiveness in improving network performance metrics, including SE and interference mitigation.

Paper Structure

This paper contains 16 sections, 8 equations, 7 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related work
  3. System model
  4. Network model
  5. Codebook and PMI selection
  6. Codebook Structure
  7. PMI selection method (Follow PMI)
  8. BS Selection of PMI
  9. Problem Definition
  10. Markov Decision Process (MDP)
  11. Advantage Actor-Critic (A2C)
  12. Cell selection for optimization
  13. Performance and analysis
  14. Simulation setup and training
  15. Results and analysis
  16. ...and 1 more sections

Figures (7)

  • Figure 1: Simulation setup and training procedure
  • Figure 2: Rewards of the training of the A2C model
  • Figure 3: CDF of the spectral efficiency of the network
  • Figure 4: Average spectral efficiency per cell
  • Figure 5: Average throughput and PRBs utilization in the network
  • ...and 2 more figures