Machine Learning in Near-Field Communication for 6G: A Survey

2025-09-20

AI Summary

To address the near-field spherical-wave propagation induced by extremely large-scale antenna arrays (ELAAs) in 6G systems, this paper systematically investigates the pivotal role of machine learning (ML) in near-field channel modeling, low-complexity beamforming, and physical-layer security. We propose the first ML-driven near-field communications framework, integrating deep learning for accurate channel estimation, reinforcement learning for dynamic beam optimization, and federated learning for privacy-preserving distributed training—achieving high accuracy, low overhead, and strong robustness. Innovatively, we establish an intelligent, data-driven channel modeling paradigm and a lightweight beam design methodology tailored for 6G. Furthermore, we present the first comprehensive technology roadmap for ML-enabled near-field communications, explicitly identifying core challenges—including data privacy and edge computational constraints—and open research problems. This work lays both theoretical foundations and practical guidelines for intelligent 6G wireless networks.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Human-in-the-loop Machine Learning
• Multiagent Systems: Multiagent Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Security and privacy of machine learning and AI applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

6G wireless communication networks are expected to use extremely large-scale antenna arrays (ELAAs) to support higher throughput, massive connectivity, and improved system performance. ELAAs would fundamentally alter wave characteristics, transforming them from plane waves into spherical waves, thereby operating in the near field. Near-field communications (NFC) offer unique advantages to enhance system performance, but also present significant challenges in channel modeling, computational complexity, and beamforming design. The use of machine learning (ML) is emerging as a powerful approach to tackle such challenges and has the capabilities to enable intelligent, secure, and efficient 6G wireless communications. In this survey, we discuss ML-driven approaches for NFC. We first outline the fundamental concepts of NFC and ML. We then discuss ML applications in channel estimation, beamforming design, and security enhancement. We also highlight key challenges (e.g., data privacy and computational overhead). Finally, we discuss open issues and future directions to emphasize the role of advanced ML techniques in near-field system design.

Problem

• Addressing near-field communication challenges in 6G networks
• Applying machine learning to NFC channel modeling and beamforming
• Overcoming computational complexity and security issues in ELAAs

Innovation

• Machine learning for near-field communication challenges
• ML applications in channel estimation and beamforming
• Advanced ML techniques for near-field system design