AI-Driven Terahertz 6G Wireless Communication Systems: From Hardware Design to Network Intelligence

Abstract

The rapid growth of data-intensive applications, such as immersive extended reality and autonomous systems, is driving wireless networks beyond the capabilities of fifth-generation (5G) systems. Sixth-generation (6G) communication, leveraging the terahertz (THz) spectrum (0.1–10 THz), promises terabit-per-second data rates, ultra-low latency, and massive device connectivity. Yet, THz systems face critical barriers including severe propagation loss, hardware constraints, and complex spectrum management. Artificial intelligence (AI) has emerged as a key enabler to address these challenges by enhancing channel modeling, adaptive beamforming, interference mitigation, and intelligent resource allocation. This review critically examines AI-driven THz 6G wireless communication systems, covering fundamentals of THz propagation, AI integration into hardware design, and network intelligence. Representative machine learning and deep learning models are analyzed, supported by schematic frameworks and comparative summaries, to highlight their applications and limitations. Emerging approaches, including federated learning, graph neural networks, explainable AI, and physics-informed modeling, are also discussed, underscoring their role in building resilient and energy-efficient networks. The paper identifies major open issues in hardware feasibility, energy consumption, data availability, and system security, before outlining future research pathways. Recommendations emphasize lightweight AI models, hybrid physics-informed frameworks, security-aware algorithms, and accelerated standardization. By bridging electrical engineering and computer science, AI-driven THz 6G networks are poised to form the backbone of future wireless ecosystems and enable transformative applications across Industry 5.0, healthcare, and immersive digital environments.