In this paper, we study a three-dimensional (3D) spectrum sharing between device-to-device (D2D) and unmanned aerial vehicles (UAVs) communications. We consider that UAVs perform spatial spectrum sensing to opportunistically access the licensed channels that are occupied by the D2D communications of ground users. The objective of the considered 3D spectrum sharing networks is to maximize the area spectral efficiency (ASE) of UAV networks while guaranteeing the required minimum ASE of D2D networks. Using the tools from machine learning, we obtain the probability of spatial false alarm and the probability of spatial missed detection at the UAV, which helps us to characterize the density of active UAVs. Then, based on the Neyman-Pearson criterion, we further derive the coverage probability of D2D and UAV communications by leveraging the tools from stochastic geometry. In addition, the ASE of the D2D and UAV networks are also obtained. Simulation results show that a decrease in the spatial spectrum sensing radius of UAVs reduces the coverage probability of UAV communications but improves the ASE of UAV networks. Furthermore, the proposed tools allow obtaining the optimal spatial spectrum sensing radius of UAVs given certain network parameters.

中文翻译：

---

混合 D2D 和无人机网络的 3D 频谱共享

在本文中，我们研究了设备到设备 (D2D) 和无人机 (UAV) 通信之间的三维 (3D) 频谱共享。我们认为无人机执行空间频谱感知以机会访问被地面用户的 D2D 通信占用的许可信道。所考虑的 3D 频谱共享网络的目标是最大化无人机网络的区域频谱效率 (ASE)，同时保证 D2D 网络所需的最小 ASE。使用机器学习的工具，我们获得了无人机的空间误报概率和空间漏检概率，这有助于我们表征活动无人机的密度。然后，基于 Neyman-Pearson 准则，我们通过利用随机几何中的工具进一步推导出 D2D 和 UAV 通信的覆盖概率。此外，还获得了 D2D 和无人机网络的 ASE。仿真结果表明，无人机空间频谱感知半径的减小降低了无人机通信的覆盖概率，但提高了无人机网络的ASE。此外，所提出的工具允许在给定某些网络参数的情况下获得无人机的最佳空间频谱感知半径。