The global evolution of wireless technologies and intelligent sensing devices are transforming the realization of smart cities. Among the myriad of use cases, there is a need to support applications whereby low-resource IoT devices need to upload their sensor data to a remote control centre by target hard deadlines; otherwise, the data becomes outdated and loses its value, for example, in emergency or industrial control scenarios. In addition, the IoT devices can be either located in remote areas with limited wireless coverage or in dense areas with relatively low quality of service. This motivates the utilization of UAVs to offload traffic from existing wireless networks by collecting data from time-constrained IoT devices with performance guarantees. To this end, we jointly optimize the trajectory of a UAV and the radio resource allocation to maximize the number of served IoT devices, where each device has its own target data upload deadline. The formulated optimization problem is shown to be mixed integer non-convex and generally NP-hard. To solve it, we first propose the high-complexity branch, reduce and bound (BRB) algorithm to find the global optimal solution for relatively small scale scenarios. Then, we develop an effective sub-optimal algorithm based on successive convex approximation in order to obtain results for larger networks. Next, we propose an extension algorithm to further minimize the UAV’s flight distance for cases where the initial and final UAV locations are known a priori. We demonstrate the favourable characteristics of the algorithms via extensive simulations and analysis as a function of various system parameters, with benchmarking against two greedy algorithms based on distance and deadline metrics.

中文翻译：

---

从时间受限的物联网设备收集数据的无人机轨迹规划

无线技术和智能传感设备的全球演进正在改变智慧城市的实现。在无数用例中，需要支持低资源物联网设备需要在目标硬期限内将其传感器数据上传到远程控制中心的应用程序；否则，数据会过时并失去其价值，例如在紧急情况或工业控制场景中。此外，物联网设备可能位于无线覆盖范围有限的偏远地区，也可能位于服务质量相对较低的密集地区。这激发了无人机的利用，通过从具有性能保证的时间受限的物联网设备收集数据来卸载现有无线网络的流量。为此，我们联合优化无人机的轨迹和无线电资源分配，以最大化服务的物联网设备的数量，其中每个设备都有自己的目标数据上传截止日期。公式化的优化问题被证明是混合整数非凸问题，并且通常是 NP-hard。为了解决这个问题，我们首先提出了高复杂度的分支、归约和限界（BRB）算法来寻找相对较小规模场景的全局最优解。然后，我们开发了一种基于连续凸逼近的有效次优算法，以获得更大网络的结果。接下来，我们提出了一种扩展算法，以在先验已知初始和最终 UAV 位置的情况下进一步最小化 UAV 的飞行距离。