The Internet of Things (IoT) possesses enormous potential for a variety of smart agriculture use cases such as pest control, soil management, and autonomous irrigation. Due to the complex terrain and cost constraints, traditional infrastructure-based ubiquitous communication network has poor scalability and cost-efficiency. In such contexts, finding alternative economic and sustainable data collection mechanism becomes paramount. In this paper, the predefined trajectory of UAV equipped with storage capability is opportunistically utilized as a delay-tolerant data transportation channel. We propose an air-ground collaborative data delivery framework and maximize the end-to-end (E2E) data transmission efficiency through jointly optimizing the terrestrial subnet transmission scheduling strategy, subnet resource allocation strategy, subnet formation strategy, and flight speed control of opportunistic UAV. On account of the heterogeneous transmission demands and the task-oriented mobility, the terrestrial IoTs actively pre-network and aggregate the environmental information towards the cluster heads with the position advantage, to improve the data uploading efficiency. We derive the closed-form subnet transmission scheduling and subnet resource allocation strategy. The subnet formation sub-problem is constructed as a coalition formation game, which can be efficiently solved by the better response method. Considering the limited sojourn time in the farmland, the opportunistic UAV dynamically adjusts the flight speed to strike a balance between the traffic distribution, data uploading capability, and data downloading capability. We derive the closed-form solution of the flight speed control strategy. Numerical simulations demonstrate that the proposed algorithm can expand the E2E data delivery volume and outperform the benchmark algorithms.

ABSTRACT In this paper, we investigate the anti‐jamming communication challenge for unmanned aerial vehicles (UAVs) in urban environments with strong jammers. Jamming power often far exceeds the UAVs' inherent anti‐jamming capability threshold, causing anti‐jamming measures to fail and even interrupt normal communication. To address this challenge, we propose an innovative strategy that leverages the natural shielding effect of urban buildings to enhance the anti‐jamming performance of UAV communication links. The core of this strategy lies in leveraging multiple UAVs working collaboratively to form an end‐to‐end anti‐jamming communication network in the urban environment. Specifically, we first introduce a UAV formation control mechanism for end‐to‐end collaboration—‘resonant motion’ transmission. Second, we propose an anti‐jamming algorithm for urban environments with strong jammers, combining ‘resonant motion’ transmission with the artificial potential field (APF) algorithm and rapidly exploring random tree star (RRT*) to develop a novel anti‐jamming path planning algorithm. Finally, we leverage prior knowledge of jammers, UAV formation and urban environment to enable UAV formation to evade obstacles and strong jammers in urban environment, find optimal communication positions and thereby build more robust communication links. The anti‐jamming strategy proposed in this paper provides a practical new approach to addressing the technical challenge of difficult UAV communication in urban environment with strong jammers. Simulation experiments demonstrate that UAVs can effectively address the challenge of UAV formation in urban environment through collaborative operations and intelligent algorithms, achieving reliable end‐to‐end transmission for UAV formation, outperforming traditional algorithms in both anti‐jamming performance and energy consumption.

Federated learning (FL), as an innovative paradigm of distributed learning, provides reliable support for the growing edge intelligence (EI). The limitations of traditional FL’s reliance on ground base stations (BSs) make the development of aerial server unmanned aerial vehicles (UAVs) inevitable, thereby developing the air-ground integrated FL (AGIFL). However, current efforts predominately focus on static offline training based on existing datasets and some efforts consider online training in dynamic sample environments, where new samples need to be fully pre-trained to determine sample quality. To this end, we study how to realize high-performance of FL in dynamic environment without training all samples. Specifically, we formulate a joint optimization problem for sample selection, UAV deployment, and resource distribution aiming to minimize the trade-off between the user energy consumption and FL performance. To address the optimization problem without explicit expression, we employ meta-learning to derive an upper bound on the gradient norm of the loss function to evaluate learning performance, and describe how time-varying small-batch ratios affect this bound. Then, we propose an optimization algorithm that ensures convergence, capitalizing on the block coordinate descent techniques. To demonstrate the efficacy of our algorithm, we conduct both extensive simulations and proof-of-concept field experiments. The findings indicate an average improvement of approximately 39% in reducing the objective value when compared to the benchmarks.

Recent advancements in unmanned aerial vehicle (UAV) technology have propelled the popularity of edge intelligence (EI) applications with deep learning in UAV swarm. Nevertheless, the high computational demands of deep neural networks (DNNs) conflict with the limited computing power and battery capacity of UAV. Furthermore, many UAV applications require real-time performance such as object detection and recognition. In this paper, we study how to achieve fast DNN inference in UAV swarm by the collaboration of multiple UAVs, and formulate the problem of minimizing the completion time of a series of arriving DNN inference tasks, under memory and energy constraints. To solve the aforementioned challenging problem with combinatorial explosion, we propose an efficient solution exploiting deep reinforcement learning (DRL) with action space simplification to find the allocation strategy of each DNN inference task within a resource-constrained UAV swarm. Simulation results validate the effectiveness of the proposed solution compared to five benchmark algorithms.

Abstract Spectrum map is a database that stores multidimensional representations of spectrum situation information. It provides support for spectrum sensing and endows wireless communication networks with intelligence. However, the ubiquitous deployment of monitoring devices leads to huge costs of operation and maintenance. It indicates that an approach is needed to reduce the number of monitoring devices, but prevent the degradation of data granularity. Therefore, this paper focuses on the accurate construction of the spectrum map. It aims to infer the fine‐grained spectrum situation of the target region based on coarse‐grained observation. In order to solve this problem, an inference framework based on deep residual network is developed in this paper. In the case of rule deployment for sensing nodes, it adopts the idea of super resolution to improve the accuracy of the spectrum map. The framework is composed of two major parts: an inference network, which generates fine‐grained spectrum maps from coarse‐grained counterparts by using feature extraction module and upsampling construction module; and a fusion network, which considers the influence of environmental factors to further improve the performance. A large number of experiments on simulated datasets verify the effectiveness of the proposed method.

Abstract This paper studies the problem of joint deployment, user association, channel, and resource allocation in unmanned aerial vehicle‐enabled access network. Since different user equipments performing different tasks and have different data rate requirements, the priority‐based traffic fairness problem is investigated. This problem, however, is a mixed integer nonlinear programming problem with NP‐hard complexity, making it challenging to be solved. To address this issue, a self‐organized and distributed framework “sense‐as‐you‐fly” based on the decomposition process, which divides the original problem into several subproblems, is proposed. Assuming without central controller, we derive the closed‐form resource allocation scheme and propose distributed many‐to‐one matching to optimize user association subproblem. Considering the coupled characteristics, the multi‐unmanned aerial vehicle deployment and channel allocation subproblems are modelled as a local altruistic game. The existence of Nash equilibrium is proved with the aid of exact potential game and efficient best response learning‐based algorithm is proposed. The original problem is finally addressed by solving the sub‐problems alternately and iteratively. Simulation results verify its effectiveness. By jointly optimizing multidimensional variables, the proposed algorithm unlocks network performance gains, especially in resource‐limited regimes.

As a way of providing users flexible computing services, networks exist that can make full use of air and ground computing resources. Such networks are called air-ground collaborative mobile edge computing (AGC-MEC) networks. AGC-MEC supports numerous emerging real-time applications for which timely computed results are critical. Researchers have developed a novel metric “age of information (AoI)” that can capture the freshness of computed results. This is the first paper to study the problem of AoI-aware scheduling forAir-groundCollaborative mobileEdge computing (i.e., IACE). So as to minimize the weighted AoI of all the terrestrial user equipments (UEs), we have jointly optimized task scheduling, computing resource allocation, and unmanned aerial vehicle (UAV) trajectory taking into account the constraints on the computing resources and the available energy of the UAV. The formulated problem, which is a challenge to solve, is a mixed-integer nonlinear programming (MINLP) problem. To obtain an effective solution, we propose an iterative algorithm based on the alternating optimization approach, which entails dividing the considered problem into three subproblems. Extensive simulations show that the proposed algorithm can achieve lower weighted AoI than five benchmark algorithms, while satisfying the resource constraints. Furthermore, simulation results demonstrate two interesting insights. First, the introduction of an aerial MEC server facilitates a flexible offloading design of the UEs which is critical to guaranteeing the freshness of computed results. Second, by optimizing the scheduling, the proposed design can unlock performance gains, especially in the resource-limited regime.

In this article, we study the problem of task selection and scheduling in unmanned aerial vehicle (UAV)-enabled multiaccess edge computing for reconnaissance (ASSUMER). Specifically, taking into account the time-varying priorities of reconnaissance tasks, we investigate how to maximize the overall reconnaissance utility by selecting an appropriate set of tasks and scheduling their execution sequence in the multiaccess edge computing server of the UAV. The ASSUMER problem is a mixed-integer nonlinear programming (MINLP) problem, which includes both integer and continuous variables and is proved to be NP-hard. To address this challenging problem, we first model the task scheduling subproblem as a single machine scheduling problem with the deterioration effect. We find out that the optimal task scheduling can be solved efficiently given any task selection variables and propose an optimal scheduling algorithm. Second, using the proposed scheduling algorithm, the ASSUMER problem is equivalent to a binary integer programming problem with respect to the task selection variable only. We prove that the objective function falls into the category of the submodular function and transform the original problem into the problem of maximizing submodular function with the energy constraint. Third, combining the proposed scheduling algorithm with submodularity, we design an effective approximation algorithm for the ASSUMER problem and prove that the algorithm has$(1 - {e^{ - 1}})/2$bicriterion approximation guarantee. Finally, simulation results show that the proposed algorithm can improve the overall reconnaissance utility and energy efficiency compared to five benchmark algorithms.