A Physical-layer Key Generation Method via Domain-adversarial Autoencoder in UAV Scenarios
L Y Zhang, Weijie Tan, Xia Wang, Shuai Peng, Xiyuan Chen, Chunguo Li
Guizhou University Guizhou Institute of Technology Southeast University
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
In physical-layer key generation (PKG), the short-term reciprocity of wireless channels is the theoretical basis. However, in dynamic unmanned aerial vehicle (UAV) scenarios, the ideal reciprocity is degraded. Rapid terminal mobility induces Doppler shifts, while frequent changes between line-of-sight and non-line-of-sight environments cause abrupt fluctuations in channel state information distributions. To address this issue, we propose a PKG method based on a domain-adversarial autoencoder (CSD-AE). The CSD-AE model integrates a convolutional neural network with Squeeze-and-Excitation channel attention mechanism, which learns and assigns weights to the feature channels extracted by the convolution, so as to extract reciprocal channel features. In addition, using a domain-adversarial network and a gradient-reversal layer forces the encoder to learn features that are less sensitive to propagation-environment changes. A complete PKG process is implemented and evaluated using UAV channel data generated by the QuaDRiGa platform. Experimental results show that the proposed method maintains high feature reciprocity over the entire 0–20 dB signal-to-noise ratio range. Even in the NLOS environment at 0 dB, it reduces the key disagreement rate to 0.007 at 5 m/s while maintaining a practical key generation rate, demonstrating its effectiveness in dynamic UAV communication scenarios.
逐年被引趋势
暂无年度引用数据
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
工程Wireless Communication Security Techniques
Wireless Signal Modulation Classification · Advanced Authentication Protocols Security