Deep Learning Based Adaptive Physical Layer Key Distribution

• Published in: 2023 International Conference on Networks, Communications and Intelligent Computing (NCIC) pp. 238 - 242
• Main Authors: Wang, Ruifei, Tang, Jie, Shi, Yuhao, Wen, Hong, Ho, Pin-Han, Chang, Shih Yu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 17.11.2023
• Subjects: Autonomous aerial vehicles; Deep learning; Encryption; Feature extraction; key distribution; Physical layer; Physical layer security; Robustness; Wireless communication
• DOI: 10.1109/NCIC61838.2023.00046

With the rapid development of modern wireless communication, communication has become more and more convenient and widely popular. However, due to the broadcast nature of the wireless channel, it is vulnerable to malicious attacks from third parties. During the establishment of UAV networks, given the limited computing power and storage resources of UAVs, traditional encryption methods may adversely affect their performance. Meanwhile, key distribution based on physical characteristics provides a new way of thinking. By utilizing the physical attributes of the channel to achieve key distribution, this method not only provides a high degree of security, but also greatly improves convenience. In addition, compared with traditional schemes, deep learning can automatically learn and integrate features at different levels, avoiding complex selection and combination of algorithms, thus possessing stronger generalization and robustness. Therefore, this paper proposes the use of deep learning techniques to extract channel features to increase the adaptability of key distribution. This approach is expected to provide strong support for the security and reliability of UAV networks.