Source-linked AI summary
Air-Ground Integrated Mobile Edge Networks: Architecture, Challenges and Opportunities
Nan Cheng, Wenchao Xu, Weisen Shi, Yi Zhou, Ning Lu, Haibo Zhou, Xuemin, Shen
TL;DR
Growing mobile traffic and computation-heavy IoT applications expose limitations in conventional cellular, cloud, and fixed edge architectures. The paper proposes AGMEN, a two-layer air-ground network using flexibly deployed UAVs for communication, caching, computing, and control, and discusses its applications, challenges, and research directions. Its supported conclusion is that coordinated UAV scheduling and air-ground cooperation can jointly optimize AGMEN functions within this proposed architecture.
Problem
Conventional cellular, cloud, and fixed edge architectures face growing traffic, latency, bandwidth, mobility, and heterogeneous IoT requirements.
Method
The paper proposes a two-layer AGMEN architecture in which UAVs assist drone-cells, edge caching, edge computing, and network control.
Results
The paper describes AGMEN components, IoT application support, benefits of UAV-assisted functions, and associated challenges and research directions.
Takeaways & Limitations
AGMEN provides a framework for coordinating UAV scheduling and air-ground cooperation across communication, caching, and computing functions.
Abstract
from arXiv · showhide
The ever-increasing mobile data demands have posed significant challenges in the current radio access networks, while the emerging computation-heavy Internet of things (IoT) applications with varied requirements demand more flexibility and resilience from the cloud/edge computing architecture. In this article, to address the issues, we propose a novel air-ground integrated mobile edge network (AGMEN), where UAVs are flexibly deployed and scheduled, and assist the communication, caching, and computing of the edge network. In specific, we present the detailed architecture of AGMEN, and investigate the benefits and application scenarios of drone-cells, and UAV-assisted edge caching and computing. Furthermore, the challenging issues in AGMEN are discussed, and potential research directions are highlighted.
I. INTRODUCTION
Rising mobile traffic, massive IoT connections, and computation-heavy applications strain fixed cellular and cloud architectures. The paper proposes AGMEN to integrate UAV-assisted densification, caching, and computing with ground networks.
- Data-intensive applications and massive IoT connections require greater communication and computation capacity, low latency, and support for many devices.
- Base-station-centric networks face traffic overload, while distant cloud computing creates backhaul burden and excessive communication delay.
- Mobile edge networks move functions and resources closer to users to provide high data rates, low delay, energy efficiency, and flexible deployment.
- Network densification, edge caching, and edge computing address communication capacity, backhaul load, and computation-intensive task offloading, respectively.
- Rigid mobile edge deployments struggle with dynamic vehicular communication and computation demands, motivating more flexible aerial support.
- AGMEN integrates UAV-assisted network densification, edge caching, and computing, while examining applications, challenges, and research directions.
II. AIR-GROUND INTEGRATED MOBILE EDGE NETWORKS
AGMEN is introduced as an architecture for air-ground cooperation, organized around multi-access RAN, edge caching, and edge computing.
- The proposed architecture facilitates air-ground cooperation through three crucial components: multi-access RAN, edge caching, and edge computing.
A. AGMEN Architecture
AGMEN combines a multi-UAV aerial network with a ground network of users, vehicles, and RAN infrastructure. UAVs provide communication, sensing, caching, computing, and coordination capabilities.
- A. AGMEN Architecture: AGMEN uses two layers: UAVs form the aerial network, while mobile users, vehicles, and RAN infrastructure form the ground network.
- A. AGMEN Architecture: UAVs equipped with sensors, communication modules, processors, and storage can operate as multifunctional network controllers.
- A. AGMEN Architecture: A2A communications let UAVs exchange sensing, control, and coordination information and form a flying ad hoc network.
- A. AGMEN Architecture: Drone-cells provide flexible access, UAVs and vehicles cache popular contents, and UAVs schedule computing tasks executed by vehicle computers.
B. Multi-Access RAN with Drone-Cells
Drone-cells extend mobile edge access through flexible UAV deployment, while UAV-assisted caching addresses mobility-related content delivery problems. These benefits are balanced by fronthaul capacity and interference constraints.
- B. Multi-Access RAN with Drone-Cells: Fixed HetNets may not adapt to dynamic future traffic, particularly from growing IoT devices and services.
- B. Multi-Access RAN with Drone-Cells: Drone-cell links can mitigate blockage effects and use mmWave and beamforming for high-speed fronthaul and radio access.
- B. Multi-Access RAN with Drone-Cells: Limited UAV transmit power constrains fronthaul capacity, making drone-cells more suitable for small-packet, low-rate IoT services.
- B. Multi-Access RAN with Drone-Cells: Multi-tier drone-cells increase coverage flexibility but introduce severe interference requiring careful management.
- B. Multi-Access RAN with Drone-Cells: Mobile users can lose access to cached contents after leaving a cell or handing over, causing extra delay and bandwidth consumption.
- B. Multi-Access RAN with Drone-Cells: UAV-assisted caching stores contents at drone-BSs or mobile devices, with drone-BSs scheduling distribution among nearby users.
D. Edge Computing for IoT in AGMEN
UAVs can form a flexible flying fog computing platform for IoT services, supporting sensing, local processing, and computation offloading across aerial and ground devices.
- D. Edge Computing for IoT in AGMEN: UAVs cooperating through IoT devices can form a flying fog computing platform for flexible and resilient IoT services.Computation-heavy tasks such as face recognition and VR can be offloaded to UAVs through A2G communications.
- D. Edge Computing for IoT in AGMEN: In crowd surveillance, UAV swarms collect high-quality videos and other sensed information at sports events, parades, and similar gatherings.
- D. Edge Computing for IoT in AGMEN: Collected data can be processed locally by UAVs or offloaded through drone-cell fronthaul or A2G communications to edge servers and mobile devices.Video face recognition may require detection, segmentation, and recognition using deep CNN-based computer vision.
- D. Edge Computing for IoT in AGMEN: For crowd sensing, UAVs schedule sensing tasks and evaluate costs through a publication/subscription mechanism across aerial and ground stations.The cyber-physical architecture separates communication-based scheduling and collection from real-world sensing by UAVs and vehicles.
III. CHALLENGES IN AGMEN
AGMEN’s three-dimensional mobility, dynamic topology, time-varying channels, and air-ground interactions complicate analysis and optimization, motivating integrated evaluation methods.
- III. CHALLENGES IN AGMEN: AGMEN’s three-dimensional mobility, dynamic topology, time-varying channel conditions, and frequent air-ground interactions make network analysis and optimization difficult.
- III. CHALLENGES IN AGMEN: Several challenging issues in AGMEN are presented together with potential solutions.
A. Network Interworking
AGMEN must interconnect heterogeneous aerial and ground elements despite multiprotocol requirements and dynamic topologies. SDN provides a management framework for coordinated control and resource allocation.
- A. Network Interworking: AGMEN interworking must address heterogeneous devices using different communication technologies and dynamic topologies that degrade communication-channel quality.Interfaces and mobility-aware schemes are needed for intra-network and air-ground communications.
- B. SDN-based Cooperative Control and Communication: Dedicated control schemes must coordinate UAV movement, charging and discharging, and communication and computation task scheduling.
- B. SDN-based Cooperative Control and Communication: SDN separates control and data planes, enabling global resource and function allocation to improve AGMEN flexibility, efficiency, interoperability, and reliability.
- B. SDN-based Cooperative Control and Communication: In the SDN framework, vehicles and UAVs act as switches and crowdsensing nodes, while base stations act as controllers making network and resource-allocation decisions.SDN flows control network behavior, and selected vehicles and UAVs can reduce control-message traffic and improve control efficiency.
C. Cognition, Prediction and Optimization of Communication Links
AGMEN introduces diverse aerial, ground, and vehicular links with distinct channel features and QoS requirements, creating a need for dynamic modeling, prediction, optimization, and integrated evaluation.
- C. Cognition, Prediction and Optimization of Communication Links: AGMEN includes U2U, U2V, and U2B links alongside traditional V2V and V2I links, each with different features and QoS requirements.UAV 3D mobility affects U2U antenna direction, while high mobility produces Doppler shifts and fading on U2V links.
- C. Cognition, Prediction and Optimization of Communication Links: U2B links require high QoS when UAVs operate as drone-BSs and support large throughput.
- D. Evaluation Methods of AGMEN: AGMEN evaluation must simulate, test, and validate a heterogeneous system combining HetNets, dynamic 3D mobility, air-ground integration, and varied applications.
- D. Evaluation Methods of AGMEN: HIL and SDN can support integrated evaluation involving multiple simulation platforms and real systems, while software-defined reconstruction reduces system complexity.Vehicle and UAV road-test data can support data analysis and simulation.
IV. OPEN RESEARCH ISSUES FOR AGMEN
AGMEN research remains at an early stage, with mobile routing identified as a key aerial-network problem and stochastic optimization tools proposed for routing configuration.
- AGMEN research is still in its infant stage, leaving many key research issues open.
- Mobile routing is a key AGMEN problem, especially for aerial networks, because UAV three-dimensional mobility increases routing-topology complexity.
- Stochastic geometry models can describe mobile routing configuration and modeling problems, while K-connected center cost functions can calculate routing constraints.
B. Stochastic Optimization of Multi-Dimension AGMEN Channel
AGMEN channel optimization must address uncertain, time-varying aerial-ground wireless links while jointly considering UAV scheduling and communication, caching, and computing performance.
- Aerial-ground cooperation creates severe wireless-channel uncertainty, requiring careful study to enhance network performance.
- GSMC modeling draws transmitters, receivers, and scatterers stochastically in three-dimensional space for AGMEN channel analysis.
- The Galerkin Projection based Pattern Downgrade method is proposed for high-dimensional reverse uncertainty quantification in AGMEN channel optimization.
- UAV mobility and mission scheduling is critical because UAV energy is constrained while UAVs provide and control multiple AGMEN functions.
- Scheduling should jointly trade off mobile-data accommodation and edge-computing efficiency across drone-cell and IoT-computing roles.
D. UAV-Assisted Data Delivery
UAVs assist AGMEN data delivery through dynamically deployed, high-quality line-of-sight links and mobility-based delivery, supporting both connectivity and delay-tolerant transfer.
- UAVs can provide high-reliability, high-rate air-to-ground links where sparse ground networks cannot establish direct communication.
- As drone base stations, UAVs enhance ground connectivity, while altitude and transmission-power adjustments trade off coverage against interference.
- UAV mobility enables delay- or disruption-tolerant delivery, with routing protocols based on predicted UAV and ground-node mobility.
- AGMEN combines efficient UAV scheduling and air-ground cooperation to jointly optimize drone-cell, edge-caching, and edge-computing performance.