Trustworthy Intelligent Networks for Low-Altitude Economy
Mengshu Song, Yijing Lin, Jiacheng Wang, Geng Sun, Dong Chen, Nan Ma, Dusit Niyato, Ping Zhang
Abstract
Intelligent networks are critical components in the low-altitude economy (LAE) to enable efficient and time-sensitive applications in many domains, such as e-commerce, transportation, logistics, and public safety. This article explores the concept of trustworthy intelligent networks for LAE, emphasizing their critical role in ensuring secure coordination among unmanned aerial vehicles (UAVs), ground stations, and users. First, we discuss trustworthy intelligent networks in terms of network trust, computing trust, and data trust, mitigating security vulnerabilities. Next, we present the architecture of trustworthy UAV-enabled semantic communication for LAE and examine the trustworthy techniques required for its implementation. To enhance data trust within this architecture, we design a federated unlearning framework that mitigates malicious data effects without retraining a semantic knowledge base from scratch. A use case and application scenarios illustrate its necessity. The simulation results demonstrate the effectiveness of federated unlearning in enhancing the trustworthiness of UAV-enabled semantic communication. Finally, this article discusses the future directions of trustworthy intelligent networks for LAE.