面向感知与AI协同任务的无人机巡检多维资源联合优化算法

李侍阳; 朱晓荣

doi:10.11999/JEIT251284

面向感知与AI协同任务的无人机巡检多维资源联合优化算法

doi: 10.11999/JEIT251284 cstr: 32379.14.JEIT251284

李侍阳,
朱晓荣^,

南京邮电大学通信与信息工程学院南京 210003

基金项目: 国家自然科学基金(No.92367102)，江苏省研究生科研与实践创新计划项目(No.KYCX22_0944)

详细信息

作者简介:
李侍阳：男，硕士生，研究方向为无人机、6G网络、多维资源调度等

朱晓荣：女，博士，教授，研究方向为6G通信系统、物联网、区块链、网络大数据等

通讯作者:
朱晓荣　xrzhu@njupt.edu.cn

中图分类号: TN929.5
计量
- 文章访问数: 2
- HTML全文浏览量: 0
- PDF下载量: 0
- 被引次数: 0
出版历程
- 修回日期: 2026-03-03
- 录用日期: 2026-03-03
- 网络出版日期: 2026-03-15

Multi-dimensional Resource Joint Optimization Algorithm for UAV Inspection of Collaborative Tasks of Perception and AI

LI Shiyang,
ZHU Xiaorong^,

School of Communication and Information Engineering, Nanjing University of Posts and Telecommunications, Nanjin 210003, China

Funds: The Natural Science Foundation of China (No.92367102), The Postgraduate Research & Practice Innovation Program of Jiangsu Province (No.KYCX22_0944)

摘要

摘要: 针对在无人机感知与故障检测并发场景下，无人机巡检过程中任务复杂，带宽、算力、功率等多维资源调度困难的问题，本文提出了一种面向感知与AI (Artificial Intelligence)协同规划的智能无人机巡检多维资源联合优化算法。首先本文提出了一种单无人机与多个计算节点之间联合协同完成多个任务的框架，在无人机与多个边缘计算节点协同工作的系统框架下，无人机在巡检点采集图像与传感器数据，并将其分批传输至多个节点进行分布式处理，完成飞行状态感知与故障检测任务，最终形成了以无人机系统能耗最小化为目标，带宽、功率、算力、节点选择、数据量和压缩率为变量的最优化问题。针对该问题，将原优化问题分解为4个子问题，分别采用双辅助混合整数线性规划(Mixed Integer Linear Programming, MILP)转化、数据驱动边界学习、基于逐次凸逼近(Successive Convex Approximation, SCA)的带宽功率联合优化和下界解析分配等方法进行求解，并通过交替优化策略实现整体优化，并进行了复杂度分析。最后仿真结果表明，与其它先进算法相比，本文提出的方法在时延、能耗、精度方面具有更优的性能。
- 无人机 /
- 智能巡检 /
- 资源调度 /
- 联合优化
Abstract: Objective With the increasing demand for aerial activities, the operational capabilities of various aircraft are gradually expanding to all airspace and multiple industries. The application scope of UAVs has covered multiple altitude layers from low to high altitudes, including micro, medium and large models, and is widely used in public safety, transportation, emergency management, logistics distribution, geographic surveying and mapping and other fields, continuously promoting the innovation and transformation of production and lifestyle. Compared with the traditional manual inspection method, UAV inspection, as an emerging business, can obtain image information that is difficult for the human eye to capture, which not only significantly reduces labor costs, but also improves the accuracy and efficiency of inspection operations. However, UAV inspection also poses new challenges to the allocation of multidimensional resources and task scheduling planning. Taking power system inspection as an example, transmission lines are exposed to the outdoors for a long time and are prone to corrosion, aging and even damage, and need to rely on regular inspections to ensure operational safety. Methods The four-stage multi-dimensional resource inspection and scheduling collaborative optimization algorithm decomposes the original optimization problem into four sub-problems based on the inspection process. After mathematical analysis of each sub-problem, a corresponding solution method is proposed. For the node selection problem, a dual-aided MILP transformation method is used; for the UAV data acquisition problem, a data-driven boundary learning method is employed; for the UAV communication resource allocation problem, a bandwidth-power joint optimization algorithm based on SCA is used; and for the node computing power allocation problem, a lower-bound analytical allocation method is employed. Finally, the original problem is solved using an alternating optimization method for the sub-problems, forming the entire algorithm. Results and Discussions Simulation results show that the proposed algorithm improves the overall energy consumption of the UAV compared to the comparative algorithms. This paper conducts simulation training on visual positioning and fault detection services, investigating the relationship between compression ratio and data volume and both. Figures 2-5 show that the fault detection accuracy is optimal when using 60% data volume and 60% compression ratio. Visual positioning accuracy is optimal when using 80% data volume and 80% compression ratio. Figure 6 shows that the proposed algorithm outperforms the comparative algorithms in terms of accuracy for AI services. As shown in Figures 7 and 8, with changes in bandwidth, computing power, and other resources, the proposed algorithm consistently outperforms the comparative algorithms in terms of energy consumption, effectively reducing overall energy consumption. Conclusions This paper proposes a multi-dimensional resource joint optimization algorithm for intelligent UAV inspection, focusing on the collaborative optimization of perception and AI. It forms an optimization problem with minimizing UAV energy consumption as the objective, and bandwidth, power, computing power, node selection, data volume, and actual compression ratio as variables. This algorithm simultaneously minimizes UAV energy consumption for both fault detection and visual localization, two AI services. Simulation results show that the algorithm can reduce the total energy consumption of the UAV and improve the accuracy of model training. This research focuses on the application scenario of single-UAV inspection; future research can further explore more complex multi-UAV collaborative inspection scenarios and incorporate more services for comprehensive study.
- Unmanned Aerial Vehicle /
- Intelligent Inspection /
- Resource Scheduling /
- Joint Optimization

HTML全文

图 1 无人机巡检业务场景示意图

下载: 全尺寸图片幻灯片

图 2 数据量对故障检测精度影响

下载: 全尺寸图片幻灯片

图 3 实际压缩比对故障检测精度影响

下载: 全尺寸图片幻灯片

图 4 数据量对视觉定位精度影响

下载: 全尺寸图片幻灯片

图 5 实际压缩比对视觉定位精度影响

下载: 全尺寸图片幻灯片

图 6 不同算法下准确率与误差距离对比

下载: 全尺寸图片幻灯片

图 7 无人机能耗与节点平均算力关系图

下载: 全尺寸图片幻灯片

图 8 无人机能耗与总带宽关系图

下载: 全尺寸图片幻灯片

参考文献(22)

[1]	谷美颖, 李航, 张家伟, 等. 基于视觉的无人机定位与导航方法研究综述[J]. 电子学报, 2025, 53(3): 651–685. doi: 10.12263/DZXB.20240699. GU Meiying, LI Hang, ZHANG Jiawei, et al. A review of vision-based UAV localization and navigation methods[J]. Acta Electronica Sinica, 2025, 53(3): 651–685. doi: 10.12263/DZXB.20240699.
[2]	YANG Yuwei, HE Minheng, LIU Juan, et al. The UAV intelligent inspection technology in the transformer substation inspection[C]. 2025 5th Power System and Green Energy Conference, Hong Kong, China, 2025: 215–219. doi: 10.1109/PSGEC66102.2025.11151059.
[3]	肖国德, 张贺. 基于深度学习的无人机输变配一体化巡检系统[J]. 自动化应用, 2024, 65(23): 4–6. doi: 10.19769/j.zdhy.2024.23.002. XIAO Guode and ZHANG He. Deep learning-based unmanned aerial vehicle integrated inspection system for power transmission transformation and distribution[J]. Automation Application, 2024, 65(23): 4–6. doi: 10.19769/j.zdhy.2024.23.002.
[4]	HE Zhenyao, XU Wei, SHEN Hong, et al. Integrated sensing and full-duplex communication: Joint transceiver beamforming and power allocation[C]. 2023 IEEE International Conference on Acoustics, Speech and Signal Processing, Rhodes Island, Greece, 2023: 1–5. doi: 10.1109/ICASSP49357.2023.10097111.
[5]	DONG Huanyu, LI Peichun, DAI Minghui, et al. Coordinated multi-point aided integrated sensing, communication and computation system: An energy efficient design[C]. 2024 IEEE/CIC International Conference on Communications in China, Hangzhou, China, 2024: 54–59. doi: 10.1109/ICCC62479.2024.10681790.
[6]	DENG Cailian, FANG Xuming, and WANG Xianbin. Integrated sensing, communication, and computation with adaptive DNN Splitting in multi-UAV networks[J]. IEEE Transactions on Wireless Communications, 2024, 23(11): 17429–17445. doi: 10.1109/TWC.2024.3453650.
[7]	ZHANG Ruizhi, ZHANG Ying, TANG Rui, et al. A joint UAV Trajectory, User association, and beamforming design strategy for multi-UAV-assisted ISAC systems[J]. IEEE Internet of Things Journal, 2024, 11(18): 29360–29374. doi: 10.1109/JIOT.2024.3430390.
[8]	ZHANG Xin, CHANG Zheng, ZHANG Guopeng, et al. Trajectory optimization and resource allocation for time minimization in the UAV-enabled MEC system[C]. 2022 IEEE Wireless Communications and Networking Conference, Austin, USA, 2022: 333–338. doi: 10.1109/WCNC51071.2022.9771719.
[9]	王怡, 覃团发, 韦睿, 等. SAG-MEC网络下支持WPT的无人机动态任务卸载与资源分配[J/OL]. 计算机工程, 1–10. https://doi.org/10.19678/j.issn.1000-3428.0070030, 2024. WANG Yi, QIN Tuanfa, WEI Rui, et al. Dynamic task unloading and resource allocation of UAVs supported by WPT in SAG-MEC network[J/OL]. Computer Engineering, 1–10. https://doi.org/10.19678/j.issn.1000-3428.0070030, 2024.
[10]	王轶宇, 钱鹏智, 张余, 等. 基于联盟博弈的无人机集群任务分配与频谱资源联合规划方法[J]. 中国电子科学研究院学报, 2024, 19(7): 647–657. doi: 10.3969/j.issn.1673-5692.2024.07.009. WANG Yiyu, QIAN Pengzhi, ZHANG Yu, et al. Task allocation and joint spectrum resource planning for UAV cluster based on alliance game theory[J]. Journal of China Academy of Electronics and Information Technology, 2024, 19(7): 647–657. doi: 10.3969/j.issn.1673-5692.2024.07.009.
[11]	BAYESSA G A, CHAI Rong, LIANG Chengchao, et al. Content fetching delay optimization-based caching and resource allocation for UAV-enabled networks[J]. IEEE Access, 2024, 12: 62429–62447. doi: 10.1109/ACCESS.2024.3395279.
[12]	GAO Yuan, DING Yu, WANG Ye, et al. Deep reinforcement learning-based trajectory optimization and resource allocation for secure UAV-enabled MEC networks[C]. IEEE Conference on Computer Communications Workshops, Vancouver, Canada, 2024: 1–5. doi: 10.1109/INFOCOMWKSHPS61880.2024.10620895.
[13]	WANG Jiawei and SUN Haifeng. Joint resource allocation and trajectory optimization for computation offloading in UAV-enabled mobile edge computing[C]. 2024 6th International Conference on Communications, Information System and Computer Engineering, Guangzhou, China, 2024: 302–307. doi: 10.1109/CISCE62493.2024.10653394.
[14]	HU Bintao, ISAAC M, AKINOLA O M, et al. Federated learning empowered resource allocation in UAV-assisted edge intelligent systems[C]. 2023 IEEE 3rd International Conference on Computer Communication and Artificial Intelligence, Taiyuan, China, 2023: 336–341. doi: 10.1109/CCAI57533.2023.10201325.
[15]	YIN Sixing and YU F R. Resource allocation and trajectory design in UAV-aided cellular networks based on multiagent reinforcement learning[J]. IEEE Internet of Things Journal, 2022, 9(4): 2933–2943. doi: 10.1109/JIOT.2021.3094651.
[16]	PENG Sicong, LI Bin, LIU Lei, et al. Trajectory design and resource allocation for multi-UAV-assisted sensing, communication, and edge computing integration[J]. IEEE Transactions on Communications, 2025, 73(4): 2847–2861. doi: 10.1109/TCOMM.2024.3478115.
[17]	AL-HOURANI A, KANDEEPAN S, and LARDNER S. Optimal LAP altitude for maximum coverage[J]. IEEE Wireless Communications Letters, 2014, 3(6): 569–572. doi: 10.1109/LWC.2014.2342736.
[18]	XIE Hao, ZHANG Tiankui, XU Xiaoxia, et al. Joint sensing, communication, and computation in UAV-assisted systems[J]. IEEE Internet of Things Journal, 2024, 11(18): 29412–29426. doi: 10.1109/JIOT.2024.3362937.
[19]	MCMAHAN B, MOORE E, RAMAGE D, et al. Communication-efficient learning of deep networks from decentralized data[C]. Proceedings of the 20th International Conference on Artificial Intelligence and Statistics, Fort Lauderdale, USA, 2017: 1273–1282.
[20]	ISMAIL M and LE Longbao. Computation offloading and resource allocation for deep neural network inference in UAV wireless networks[C]. IEEE International Conference on Communications, Montreal, Canada, 2025: 832–837. doi: 10.1109/ICC52391.2025.11161364.
[21]	WANG Die, JIA Yunjian, LIANG Liang, et al. Resource allocation in blockchain integration of UAV-enabled MEC networks: A stackelberg differential game approach[J]. IEEE Transactions on Services Computing, 2024, 17(6): 4197–4210. doi: 10.1109/TSC.2024.3418330.
[22]	CAO Xiaolan, YAN Feng, XIA Weiwei, et al. DDQN based adaptive multi-channel MAC protocol for UAV ad-hoc networks[C]. 2024 10th International Conference on Computer and Communications, Chengdu, China, 2024: 1680–1685. doi: 10.1109/ICCC62609.2024.10942236.