DNN Adaptive Partitioning Strategy for Heterogeneous Online Inspection Systems of Substations

被引:0
|
作者
Fu, Qincui [1 ]
Deng, Fangming [1 ]
Xue, Xianfa [1 ]
Zeng, Jianjun [1 ]
Wei, Baoquan [1 ]
机构
[1] East China Jiaotong Univ, Sch Elect & Automat Engn, Nanchang 330013, Peoples R China
基金
中国国家自然科学基金;
关键词
DNN; UAV; electricity inspection; EDGE;
D O I
10.3390/electronics13173383
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
With the explosive development of power edge equipment and the continuous improvement in power inspection performance, the requirements of substations and terminal equipment, such as drones with limited resources, cannot meet the strict delay and energy consumption requirements. This paper proposes an adaptive partitioning strategy for heterogeneous substation inspection systems. First, a layer delay prediction model and layer energy consumption prediction model are established on each heterogeneous node, and nonlinear characteristics related to delay and energy consumption are trained. On this basis, a deep neural network (DNN) hybrid partitioning strategy is proposed. The DNN task is divided into synchronous cooperative reasoning between terminal devices and multi-heterogeneous edge nodes. The experimental results show that the average absolute percentage error (MAPE) of the delay model was reduced by 31.49% on average. On drones and mobile edge nodes, the energy consumption model MAPE reduced the average by 21.92%, and the DNN end-to-end latency was reduced by 31.48%. The total cost of the system was reduced and the efficiency of UAV inspection was improved.
引用
收藏
页数:17
相关论文
共 50 条
  • [1] Adaptive partitioning and efficient scheduling for distributed DNN training in heterogeneous IoT environment
    Huang, Binbin
    Huang, Xunqing
    Liu, Xiao
    Ding, Chuntao
    Yin, Yuyu
    Deng, Shuiguang
    [J]. COMPUTER COMMUNICATIONS, 2024, 215 : 169 - 179
  • [2] CoEdge: Cooperative DNN Inference With Adaptive Workload Partitioning Over Heterogeneous Edge Devices
    Zeng, Liekang
    Chen, Xu
    Zhou, Zhi
    Yang, Lei
    Zhang, Junshan
    [J]. IEEE-ACM TRANSACTIONS ON NETWORKING, 2021, 29 (02) : 595 - 608
  • [3] RoaD-RuNNer: Collaborative DNN partitioning and offloading on heterogeneous edge systems
    Kakolyris, Andreas Kosmas
    Katsaragakis, Manolis
    Masouros, Dimosthenis
    Soudris, Dimitrios
    [J]. 2023 DESIGN, AUTOMATION & TEST IN EUROPE CONFERENCE & EXHIBITION, DATE, 2023,
  • [4] Distributed Inference Acceleration with Adaptive DNN Partitioning and Offloading
    Mohammed, Thaha
    Joe-Wong, Carlee
    Babbar, Rohit
    Di Francesco, Mario
    [J]. IEEE INFOCOM 2020 - IEEE CONFERENCE ON COMPUTER COMMUNICATIONS, 2020, : 854 - 863
  • [5] An Inspection-maintenance Strategy for Heterogeneous Systems with Measurable Degradation
    Ye, Z. S.
    Zhang, M. M.
    Xiao, X.
    [J]. 2013 IEEE INTERNATIONAL CONFERENCE ON INDUSTRIAL ENGINEERING AND ENGINEERING MANAGEMENT (IEEM 2013), 2013, : 1432 - 1436
  • [6] A partitioning strategy of multiple model adaptive control for interval systems
    Choi, SW
    Cho, TS
    Kim, YC
    [J]. ADVANCES IN CONTROL EDUCATION 2000, 2001, : 349 - 354
  • [7] Online horizontal partitioning of heterogeneous data
    Herrmann, Kai
    Voigt, Hannes
    Lehner, Wolfgang
    [J]. IT-INFORMATION TECHNOLOGY, 2014, 56 (01): : 4 - 12
  • [8] Adaptive Informative Sampling with Environment Partitioning for Heterogeneous Multi-Robot Systems
    Shi, Yunfei
    Wang, Ning
    Zheng, Jianmin
    Zhang, Yang
    Yi, Sha
    Luo, Wenhao
    Sycara, Katia
    [J]. 2020 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2020, : 11718 - 11723
  • [9] Privacy-Aware Edge Computing Based on Adaptive DNN Partitioning
    Shi, Chengshuai
    Chen, Lixing
    Shen, Cong
    Song, Linqi
    Xu, Jie
    [J]. 2019 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM), 2019,
  • [10] Towards Resource-aware DNN Partitioning for Edge Devices with Heterogeneous Resources
    Zawish, Muhammad
    Abraham, Lizy
    Dev, Kapal
    Davy, Steven
    [J]. 2022 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM 2022), 2022, : 5649 - 5655