Deep Learning-Driven Intelligent THz Communications for 6G Urban Air Mobility: Channel Prediction and URLLC Assurance

5月 1, 2026·
Guixiang Zhu
,
Yuqin Ou
,
Qi Sun
,
Jiaqi Miao
,
Dasong Sun
,
Beiping Zhou
· 0 分钟阅读时长
URL DOI
摘要
With the rapid evolution of sixth-generation (6G) mobile communications, urban air mobility (UAM) networks have emerged as a key pillar to achieve the vision of ubiquitous space-air-ground connectivity. Although the terahertz (THz) frequency band is regarded as a critical technological enabler for supporting real-time, high-bandwidth UAM services due to its abundant spectrum resources, THz waves suffer severe propagation attenuation and are highly sensitive to dynamic fluctuations in atmospheric conditions. In this paper, we propose an intelligent THz communication system model tailored for 6G UAM networks. By incorporating a data-driven machine learning (ML) framework, the proposed model enhances the system’s adaptability to dynamic environments through the deep integration of physical propagation priors with intelligent algorithms. Based on the established system model, a deep learning-driven channel prediction and ultra-reliable low-latency communications (URLLC) assurance mechanism is introduced. Using the powerful nonlinear feature extraction capabilities of deep neural networks (DNNs), the system can identify latent atmospheric fading patterns from complex THz propagation paths, thereby providing proactive channel state information (CSI) for unmanned aerial vehicle (UAV) nodes. This predictive approach effectively mitigates the latency bottlenecks inherent in traditional reactive methods when dealing with rapidly changing spatiotemporal features. Extensive numerical simulations demonstrate that the proposed intelligent model significantly outperforms traditional benchmark schemes in terms of prediction accuracy and system throughput under extreme attenuation and highdynamic scenarios. Finally, the results confirm the significant advantages of this scheme in maintaining ultra-low end-to-end response latency, providing a robust solution for optimizing THz communication performance in complex maritime or urban environments.
类型
出版物
2026 6th International Conference on Computer Communication and Artificial Intelligence (CCAI)
publications