ABSTRACT
Roadside traffic monitoring is increasingly performed by deploying roadside high-resolution video cameras and then running Computer Vision (CV) models on the video data. Since computer vision models are compute-intensive as they utilize Deep Neural Networks (DNNs), the data is usually sent to one or more edge servers located adjacent to mobile base stations. Recent techniques propose running CV models on tiles of videos separately to detect and track small objects. Several CV models exist, each with different requirements of compute and memory. Since more compute and memory-intensive CV models provide higher accuracy, a key challenge of such techniques is to determine which vision model should be used on which tile. This becomes even more challenging if multiple videos are processed by the same edge server. In this paper, we first formulate this problem of model selection on edge devices as an Integer Linear Programming (ILP) problem, and then propose a heuristic to solve it. Our experiments show that it is quite effective in practice.
- [n.d.]. Optimization with PuLP. https://coin-or.github.io/pulp/Google Scholar
- [n.d.]. Smart Camera Video Analytics; Meraki Cameras. https://meraki.cisco.com/products/smart-cameras/video-analytics/Google Scholar
- Romil Bhardwaj, Zhengxu Xia, Ganesh Ananthanarayanan, Junchen Jiang, Nikolaos Karianakis, Yuanchao Shu, Kevin Hsieh, Victor Bahl, and Ion Stoica. 2022. Ekya: Continuous learning of video analytics models on edge compute servers. In USENIX NSDI. 119–135.Google Scholar
- Kuntai Du, Ahsan Pervaiz, Xin Yuan, Aakanksha Chowdhery, Qizheng Zhang, Henry Hoffmann, and Junchen Jiang. 2020. Server-driven video streaming for deep learning inference. In ACM SIGCOMM. 557–570.Google Scholar
- G. Fabbri, C. Boccaletti, L. Anniballi, M. London, F. Calenne, F. M. F. Mascioli, and A. J. M. Cardoso. 2013. Adding ICT to Transportation Systems: Monitoring and Management System for the Electric Vehicle Fleet on the Island of Ventotene. In International Journal of Computer Theory and Engineering, Vol. 5. 268–272.Google Scholar
- Kevin Hsieh, Ganesh Ananthanarayanan, Peter Bodik, Shivaram Venkataraman, Paramvir Bahl, Matthai Philipose, Phillip B. Gibbons, and Onur Mutlu. 2018. Focus: Querying Large Video Datasets with Low Latency and Low Cost. In OSDI. 269–286.Google Scholar
- Chien-Chun Hung, Ganesh Ananthanarayanan, Peter Bodik, Leana Golubchik, Minlan Yu, Paramvir Bahl, and Matthai Philipose. 2018. VideoEdge: Processing Camera Streams using Hierarchical Clusters. In IEEE/ACM SEC. 115–131. https://doi.org/10.1109/SEC.2018.00016Google ScholarCross Ref
- Faraz Islam and Ali Faraz Syed. 2014. PLC Based Intelligent Toll Road Traffic Control Using. In International Journal of Computer Theory and Engineering, Vol. 6. 353–356.Google Scholar
- Junchen Jiang, Ganesh Ananthanarayanan, Peter Bodik, Siddhartha Sen, and Ion Stoica. 2018. Chameleon: Scalable Adaptation of Video Analytics. In SIGCOMM (Budapest, Hungary). 253–266. https://doi.org/10.1145/3230543.3230574Google ScholarDigital Library
- Ensheng Liu, Liping Zheng, Qiang He, Benzhu Xu, and Gaofeng Zhang. 2021. Criticality-awareness edge user allocation for public safety. IEEE TSC (2021).Google ScholarCross Ref
- Arvind Narayanan, Eman Ramadan, Rishabh Mehta, Xinyue Hu, Qingxu Liu, Rostand A.K. Fezeu, Udhaya Kumar Dayalan, Saurabh Verma, Peiqi Ji, Tao Li, Feng Qian, and Zhi-Li Zhang. 2021. Lumos5G Dataset. https://doi.org/10.1145/3419394.3423629Google ScholarDigital Library
- Arvind Narayanan, Xumiao Zhang, Ruiyang Zhu, Ahmad Hassan, Shuowei Jin, Xiao Zhu, Xiaoxuan Zhang, Denis Rybkin, Zhengxuan Yang, Zhuoqing Morley Mao, 2021. A variegated look at 5G in the wild: performance, power, and QoE implications. In ACM SIGCOMM. 610–625.Google Scholar
- F. Ozge Unel, Burak O. Ozkalayci, and Cevahir Cigla. 2019. The Power of Tiling for Small Object Detection. In IEEE CVPR Workshops. 582–591.Google Scholar
- Subrat Prasad Panda, Kaustabha Ray, and Ansuman Banerjee. 2020. Dynamic edge user allocation with user specified QoS preferences. In ICSOC. 187–197.Google Scholar
- Sohee Park, Arani Bhattacharya, Zhibo Yang, Samir R. Das, and Dimitris Samaras. 2021. Mosaic: Advancing User Quality of Experience in 360-Degree Video Streaming With Machine Learning. IEEE TNSM, 18(1) (2021), 1000–1015. https://doi.org/10.1109/TNSM.2021.3053183Google ScholarDigital Library
- Q. Peng, Y. Xia, Z. Feng, J. Lee, C. Wu, X. Luo, W. Zheng, S. Pang, H. Liu, Y. Qin, and P. Chen. 2019. Mobility-Aware and Migration-Enabled Online Edge User Allocation in Mobile Edge Computing. In ICWS. 91–98.Google Scholar
- Feng Qian, Bo Han, Qingyang Xiao, and Vijay Gopalakrishnan. 2018. Flare: Practical Viewport-Adaptive 360-Degree Video Streaming for Mobile Devices. In MobiCom (New Delhi, India). 99–114. https://doi.org/10.1145/3241539.3241565Google ScholarDigital Library
- Chaogang Tang, Chunsheng Zhu, Ning Zhang, Mohsen Guizani, and Joel J. P. C. Rodrigues. 2022. SDN-Assisted Mobile Edge Computing for Collaborative Computation Offloading in Industrial Internet of Things. IEEE Internet of Things Journal 9, 23 (2022), 24253–24263. https://doi.org/10.1109/JIOT.2022.3190281Google ScholarCross Ref
- S Vivienne, B Madhukar, and JS Gary. [n.d.]. High Efficiency Video Coding (HEVC): Algorithms and Architectures, 2014.Google Scholar
- Can Wang, Sheng Zhang, Yu Chen, Zhuzhong Qian, Jie Wu, and Mingjun Xiao. 2020. Joint Configuration Adaptation and Bandwidth Allocation for Edge-based Real-time Video Analytics. In IEEE INFOCOM. 257–266. https://doi.org/10.1109/INFOCOM41043.2020.9155524Google ScholarDigital Library
- Jiadai Wang, Lei Zhao, Jiajia Liu, and Nei Kato. 2019. Smart resource allocation for mobile edge computing: A deep reinforcement learning approach. IEEE TETC 9, 3 (2019), 1529–1541.Google Scholar
- Robert J Wang, Xiang Li, and Charles X Ling. 2018. Pelee: A real-time object detection system on mobile devices. Advances in neural information processing systems 31 (2018).Google Scholar
- Xu Wang, Zheng Yang, Jiahang Wu, Yi Zhao, and Zimu Zhou. 2021. EdgeDuet: Tiling Small Object Detection for Edge Assisted Autonomous Mobile Vision. In IEEE INFOCOM. 1–10. https://doi.org/10.1109/INFOCOM42981.2021.9488843Google ScholarDigital Library
- Longyin Wen, Dawei Du, Zhaowei Cai, Zhen Lei, Ming-Ching Chang, Honggang Qi, Jongwoo Lim, Ming-Hsuan Yang, and Siwei Lyu. 2020. UA-DETRAC: A New Benchmark and Protocol for Multi-Object Detection and Tracking. Computer Vision and Image Understanding (2020).Google Scholar
- Tian-Hao Wu, Tong-Wen Wang, and Ya-Qi Liu. 2021. Real-Time Vehicle and Distance Detection Based on Improved Yolo v5 Network. In 2021 3rd World Symposium on Artificial Intelligence (WSAI). 24–28. https://doi.org/10.1109/WSAI51899.2021.9486316Google ScholarCross Ref
- Haoyu Zhang, Ganesh Ananthanarayanan, Peter Bodik, Matthai Philipose, Paramvir Bahl, and Michael J. Freedman. 2017. Live Video Analytics at Scale with Approximation and Delay-Tolerance. In NSDI. Boston, MA, 377–392.Google Scholar
- Wuyang Zhang, Sugang Li, Luyang Liu, Zhenhua Jia, Yanyong Zhang, and Dipankar Raychaudhuri. 2019. Hetero-Edge: Orchestration of Real-time Vision Applications on Heterogeneous Edge Clouds. In IEEE INFOCOM. 1270–1278. https://doi.org/10.1109/INFOCOM.2019.8737478Google ScholarDigital Library
- Chao Zhou, Mengbai Xiao, and Yao Liu. 2018. ClusTile: Toward Minimizing Bandwidth in 360-degree Video Streaming. In IEEE INFOCOM. 962–970. https://doi.org/10.1109/INFOCOM.2018.8486282Google ScholarDigital Library
Index Terms
- A Framework for Tile Processing on Edge Servers for Roadside Traffic Surveillance
Recommendations
Increasing Traffic Safety with Real-Time Edge Analytics and 5G
EdgeSys '21: Proceedings of the 4th International Workshop on Edge Systems, Analytics and NetworkingDespite advances in vehicle technology and road modernization, traffic accidents are a huge global issue, causing deaths and injuries, especially among pedestrians and cyclists. This often happens due to pedestrians and cyclists in drivers' blind spots ...
A Heuristic Ceiling Point Algorithm for General Integer Linear Programming
This paper first examines the role of ceiling points in solving a pure, general integer linear programming problem P. Several kinds of ceiling points are defined and analyzed and one kind called "feasible 1-ceiling points" proves to be of special ...
Deep Reinforcement Learning for Load Balancing of Edge Servers in IoV
AbstractIn recent years, the use of edge computing to solve the problem of limited resources in the IoV has attracted more and more attention. Vehicles can upload tasks to the edge servers within their radio range for computing cooperation and offloading. ...
Comments