Digital Twin-Enabled Domain Adaptation for Zero-Touch UAV Networks: Survey and Challenges

Abstract

Verbatim from arXiv. Not paraphrased, not summarized.

In existing wireless networks, the control programs have been designed manually and for certain predefined scenarios. This process is complicated and error-prone, and the resulting control programs are not resilient to disruptive changes. Data-driven control based on Artificial Intelligence and Machine Learning (AI/ML) has been envisioned as a key technique to automate the modeling, optimization and control of complex wireless systems. However, existing AI/ML techniques rely on sufficient well-labeled data and may suffer from slow convergence and poor generalizability. In this article, focusing on digital twin-assisted wireless unmanned aerial vehicle (UAV) systems, we provide a survey of emerging techniques that can enable fast-converging data-driven control of wireless systems with enhanced generalization capability to new environments. These include SLAM-based sensing and network softwarization for digital twin construction, robust reinforcement learning and system identification for domain adaptation, and testing facility sharing and federation. The corresponding research opportunities are also discussed.

Authors

• Maxwell McManus Zhaoxi
• Yuqing Cui Zhaoxi
• Josh Zhaoxi
• Zhang
• Jiangqi Hu
• Sabarish Krishna Moorthy
• Zhangyu Guan
• Nicholas Mastronarde
• Elizabeth Serena Bentley
• Michael Medley

Keywords

• cs.NI
• eess.SY