Performance Evaluation of AI/ML Model to Enhance Beam Management in 5G-Advanced System

Document Type

Conference Proceeding

Publication Date

1-30-2024

Department

Department of Physics

Abstract

In fifth generation (5G) wireless communication systems, the millimeter wave (mmWave) is identified as a significant frequency band, offering extensive spectrum resources. Despite its broad bandwidth, mmWave experiences notable path loss. This challenge can be addressed using massive multiple-input multiple-output (MIMO) technology with beamforming. However, the extensive number of beams generated to counteract path loss necessitates excessive beam quality measurements and reports to ensure users connect to the optimal beam(s). Leveraging artificial intelligence (AI) and machine learning (ML) can enhance beam management. These technologies can predict the best beams, reducing the number for measurements and reports while keeping the same performance. This paper presents an evaluation of AI/ML approaches for determining the best beams in both spatial and temporal domains. Our findings highlight the performance of AI/ML models compared to traditional methods, the generalization capabilities of AI/ML models, and their resilience against measurement and quantization errors.

Publication Title

2023 Fourteenth International Conference on Mobile Computing and Ubiquitous Network (ICMU)

Share

COinS