The Unmanned Aerial Vehicle (UAV) swarm networks will play a crucial role in the B5G/6G network thanks to its appealing features, such as wide coverage and on-demand deployment. Emergency communication (EC) is essential to promptly inform UAVs of potential danger to avoid accidents, whereas the conventional communication-only feedback-based methods, which separate the digital and physical identities (DPI), bring intolerable latency and disturb the unintended receivers. In this paper, we present a novel DPI-Mapping solution to match the identities (IDs) of UAVs from dual domains for visual networking, which is the first solution that enables UAVs to communicate promptly with what they see without the tedious exchange of beacons. The IDs are distinguished dynamically by defining feature similarity, and the asymmetric IDs from different domains are matched via the proposed bio-inspired matching algorithm. We also consider Kalman filtering to combine the IDs and predict the states for accurate mapping. Experiment results show that the DPI-Mapping reduces individual inaccuracy of features and significantly outperforms the conventional broadcast-based and feedback-based methods in EC latency. Furthermore, it also reduces the disturbing messages without sacrificing the hit rate.

翻译：无人机群网络凭借其广覆盖和按需部署等优势，将在B5G/6G网络中发挥关键作用。应急通信对及时告知无人机潜在危险以避免事故至关重要，而传统的仅基于通信反馈的方法将数字身份与物理身份分离，会带来难以忍受的延迟并干扰非目标接收者。本文提出一种新颖的DPI-Mapping解决方案，用于匹配双域中无人机的身份标识，这是首次使无人机能够无需繁琐的信标交换即可通过视觉所见实现即时通信。通过定义特征相似度动态区分身份标识，并采用所提出的仿生匹配算法匹配来自不同域的非对称身份标识。我们还结合卡尔曼滤波来融合身份标识并预测状态以实现精准映射。实验结果表明，DPI-Mapping降低了特征的个体不准确性，并在应急通信延迟方面显著优于传统的广播和反馈方法。此外，它在不牺牲命中率的情况下减少了干扰消息。