Supporting reliable and seamless connectivity for aerial users, such as Unmanned Aerial Vehicles (UAVs), is one of the key challenges for the next-generation cellular networks. To tackle this challenge, we propose a joint beamforming, user association, and UAV-height control framework for cellular-connected multi-UAV communications. Our objective is to maximize the minimum achievable rate for UAVs subject to co-existing terrestrial users’ rate constraints. A hierarchical bi-layer iterative algorithm is devised to solve the problem. By using the projection gradient method in inner layer iterations and the geometric program modeling plus the convex-concave procedure in outer layer iterations, our proposed algorithm is proved to converge to a local optimum. We also examine our proposed algorithm under the practical condition where channel estimation is not perfect. Numerical results show that our proposed joint beamforming, user association, and UAV-height control framework outperforms the conventional nearest UAV association method in terms of UAVs’ minimum achievable rate for both perfect and imperfect channel estimation cases. We also observe different UAVs’ heights (i.e., between 100m and 300m) do not affect the UAVs’ achievable rates. For the case with moving UAV, we also study the trade-off between UAVs’ minimum achievable rate and the frequency of updating optimization variables.

中文翻译：

---

蜂窝无人机通信的联合波束成形、用户关联和高度控制

支持无人机（UAV）等空中用户的可靠无缝连接是下一代蜂窝网络的主要挑战之一。为了应对这一挑战，我们为蜂窝连接的多无人机通信提出了联合波束成形、用户关联和无人机高度控制框架。我们的目标是在共存地面用户的速率限制下最大化无人机的最小可实现速率。设计了一种分层双层迭代算法来解决该问题。通过在内层迭代中使用投影梯度法和在外层迭代中使用几何程序建模和凸凹过程，证明我们提出的算法收敛到局部最优。我们还在信道估计不完美的实际条件下检查我们提出的算法。数值结果表明，我们提出的联合波束成形、用户关联和无人机高度控制框架在完美和不完美信道估计情况下无人机的最小可实现速率方面优于传统的最近无人机关联方法。我们还观察到不同无人机的高度（即在 100m 和 300m 之间）不会影响无人机的可实现速率。对于移动无人机的情况，我们还研究了无人机的最小可达速率与更新优化变量的频率之间的权衡。在完美和不完美的信道估计情况下，无人机高度控制框架在无人机的最小可实现速率方面优于传统的最近无人机关联方法。我们还观察到不同无人机的高度（即在 100m 和 300m 之间）不会影响无人机的可实现速率。对于移动无人机的情况，我们还研究了无人机的最小可达速率与更新优化变量的频率之间的权衡。在完美和不完美的信道估计情况下，无人机高度控制框架在无人机的最小可实现速率方面优于传统的最近无人机关联方法。我们还观察到不同无人机的高度（即在 100m 和 300m 之间）不会影响无人机的可实现速率。对于移动无人机的情况，我们还研究了无人机的最小可达速率与更新优化变量的频率之间的权衡。