Adaptive Edge-Cloud Data Scheduling forUrban Infrastructure Inspection Systems
DOI:
https://doi.org/10.71465/jgs84Keywords:
edge–cloud computing, stream scheduling, infrastructure inspection, Lyapunov optimization, cellular uplinks, fault toleranceAbstract
Vehicle- and drone-mounted inspection systems for urban bridges and roads generate image streams whose value depends jointly on analytic quality and timeliness, yet these streams must leave the field through cellular uplinks whose capacity fluctuates by an order of magnitude and drops to zero for tens of seconds. This paper presents ACES, an adaptive compression and edge-cloud scheduling method that decides, for every captured frame, which version to upload (resolution and JPEG quality), whether its defect analysis runs on the edge node or in the cloud, and in which order frames are transmitted and processed. ACES derives a per-frame decision rule from Lyapunov drift-plus-penalty optimization over three coupled backlogs (uplink, edge CPU, and cloud), so that backlogs act as prices for bytes and compute; it adds damped backpressure migration of uncommitted frames, an urgency-tiered preemptive order, and value-aware buffering with write-ahead-log recovery. We prove an [O(1/V),O(V)] utility-backlog trade-off for the core rule. The evaluation replays real data: rate-quality-compute profiles measured on 118 pixel-annotated urban road frames at 24 encoding versions, and five real cellular uplink traces (0.8–19.3 Mb/s, outages up to 78 s), in 1,480 trace-driven runs. ACES attains a normalized inspection utility of 0.980 versus 0.948 for the strongest baseline (paired gain +0.031, 95% CI [0.023,0.040], 24/25 wins), improves it by up to 5 points on congested low-capacity uplinks, delivers 96.8% of critical alerts within 4 s, stays ahead under edge crashes, bandwidth-estimation errors and a shared failing cloud, and spends 0.3 ms of computation per 0.2-s scheduling epoch.
References
Spencer Jr., B.F., Hoskere, V., Narazaki, Y.: Advances in computer vision-based civil infrastructure inspection and monitoring. Engineering 5(2), 199–222 (2019)
Koch, C., Georgieva, K., Kasireddy, V., Akinci, B., Fieguth, P.: A review on computer vision based defect detection and condition assessment of concrete and asphalt civil infrastructure. Advanced Engineering Informatics 29(2), 196–210 (2015)
Shi, W., Cao, J., Zhang, Q., Li, Y., Xu, L.: Edge computing: vision and challenges. IEEE Internet of Things Journal 3(5), 637–646 (2016)
Winstein, K., Sivaraman, A., Balakrishnan, H.: Stochastic forecasts achieve high throughput and low delay over cellular networks. In: Proc. USENIX NSDI, pp. 459–471 (2013)
Amazon Web Services: Manage data streams on Greengrass core devices (stream manager). AWS IoT Greengrass V2 Developer Guide, https://docs.aws.amazon.com/greengrass/v2/developerguide/manage-data-streams.html, last accessed October 2026
Fall, K.: A delay-tolerant network architecture for challenged internets. In: Proc. ACM SIGCOMM, pp. 27–34 (2003)
Zhang, B., Jin, X., Ratnasamy, S., Wawrzynek, J., Lee, E.A.: AWStream: adaptive wide-area streaming analytics. In: Proc. ACM SIGCOMM, pp. 236–252 (2018)
Jiang, J., Ananthanarayanan, G., Bodik, P., Sen, S., Stoica, I.: Chameleon: scalable adaptation of video analytics. In: Proc. ACM SIGCOMM (2018)
Zhang, S., Wang, C., Jin, Y., Wu, J., Qian, Z., Xiao, M., Lu, S.: Adaptive configuration selection and bandwidth allocation for edge-based video analytics. IEEE/ACM Transactions on Networking 30(1), 285–298 (2022)
Mao, Y., Zhang, J., Letaief, K.B.: Dynamic computation offloading for mobile-edge computing with energy harvesting devices. IEEE Journal on Selected Areas in Communications 34(12), 3590–3605 (2016)
Tran, T.X., Pompili, D.: Joint task offloading and resource allocation for multi-server mobile-edge computing networks. IEEE Transactions on Vehicular Technology 68(1), 856–868 (2019)
Braun, T.D., Siegel, H.J., Beck, N., Bölöni, L.L., Maheswaran, M., Reuther, A.I., Robertson, J.P., Theys, M.D., Yao, B., Hensgen, D., Freund, R.F.: A comparison of eleven static heuristics for mapping a class of independent tasks onto heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing 61(6), 810–837 (2001)
Shi, Y., Cui, L., Qi, Z., Meng, F., Chen, Z.: Automatic road crack detection using random structured forests. IEEE Transactions on Intelligent Transportation Systems 17(12), 3434–3445 (2016)
Netravali, R., Sivaraman, A., Das, S., Goyal, A., Winstein, K., Mickens, J., Balakrishnan, H.: Mahimahi: accurate record-and-replay for HTTP. In: Proc. USENIX ATC, pp. 417–429 (2015)
Sato, Y., Nakajima, S., Shiraga, N., Atsumi, H., Yoshida, S., Koller, T., Gerig, G., Kikinis, R.: Three-dimensional multi-scale line filter for segmentation and visualization of curvilinear structures in medical images. Medical Image Analysis 2(2), 143–168 (1998)
Wang, S., Urgaonkar, R., Zafer, M., He, T., Chan, K., Leung, K.K.: Dynamic service migration in mobile edge computing based on Markov decision process. IEEE/ACM Transactions on Networking 27(3), 1272–1288 (2019)
Wang, X., Shen, M.C., Yang, K.: On-edge high-throughput collaborative inference for real-time video analytics. IEEE Internet of Things Journal 11(20), 33097–33109 (2024). https://doi.org/10.1109/JIOT.2024.3424235
Cao, Z., Cheng, Y., Zhou, Z., Chen, Y., Hu, Y., Lu, A., Liu, J., Li, Z.: Edge-cloud collaborated object detection via bandwidth adaptive difficult-case discriminator. IEEE Transactions on Mobile Computing 24(2), 1181–1196 (2025). https://doi.org/10.1109/TMC.2024.3474743
Bai, T., Hou, B., Wang, Z., Liu, D., Nallanathan, A.: Content-aware joint knob configuration and resource allocation for edge video analytics. IEEE Transactions on Mobile Computing 24(6), 5536–5550 (2025). https://doi.org/10.1109/TMC.2025.3533596
Zhang, H., Ananthanarayanan, G., Bodik, P., Philipose, M., Bahl, P., Freedman, M.J.: Live video analytics at scale with approximation and delay-tolerance. In: Proc. USENIX NSDI, pp. 377–392 (2017)
Qian, B., Xuan, Y., Wu, D., Wen, Z., Yang, R., He, S., Chen, J., Ranjan, R.: Edge-cloud collaborative streaming video analytics with multi-agent deep reinforcement learning. IEEE Network 39(1), 165–173 (2025). https://doi.org/10.1109/MNET.2024.3398724
Liang, Y., Gao, G.: ASDMR: adaptive scheduling of DNN models and resources for collaborative edge-cloud video analytics. In: Proc. IEEE 31st International Conference on Parallel and Distributed Systems (ICPADS), pp. 1–10 (2025). https://doi.org/10.1109/ICPADS67057.2025.11323168
Mohan, C., Haderle, D., Lindsay, B., Pirahesh, H., Schwarz, P.: ARIES: a transaction recovery method supporting fine-granularity locking and partial rollbacks using write-ahead logging. ACM Transactions on Database Systems 17(1), 94–162 (1992)
Dean, J., Barroso, L.A.: The tail at scale. Communications of the ACM 56(2), 74–80 (2013)
Liu, C.L., Layland, J.W.: Scheduling algorithms for multiprogramming in a hard-real-time environment. Journal of the ACM 20(1), 46–61 (1973)
Vepsalainen, A.P.J., Morton, T.E.: Priority rules for job shops with weighted tardiness costs. Management Science 33(8), 1035–1047 (1987)
Neely, M.J.: Stochastic Network Optimization with Application to Communication and Queueing Systems. Morgan & Claypool (2010)
Wallace, G.K.: The JPEG still picture compression standard. IEEE Transactions on Consumer Electronics 38(1), xviii–xxxiv (1992)
Shang, W., Teng, D., Chen, T., Zhang, F., Ding, J.: Neuro-Elastic: a unified framework for hardware-aware adaptive quantization and dynamic sparsity in real-time edge intent prediction. IEEE Access 14, 111168–111187 (2026). https://doi.org/10.1109/ACCESS.2026.3715276
Amazon Web Services: What is AWS Panorama? AWS Panorama Developer Guide, https://docs.aws.amazon.com/panorama/latest/dev/, last accessed October 2026
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Dai Teng (Author)
This work is licensed under a Creative Commons Attribution 4.0 International License.
All articles published in the Journal of Geospace Science (JOGSS) are licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). This license allows others to share, copy, distribute, and adapt the work, provided proper attribution is given to the original author(s) and source.