Article
Chemistry, Multidisciplinary
Zonglei Lyu, Jia Luo
Summary: This article proposes a real-time object detection solution based on an edge cloud system for airport apron operation surveillance video. By utilizing a lightweight detection model and an edge video detection acceleration strategy, this solution achieves reliable and real-time monitoring on edge devices.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Yung-Yao Chen, Yu-Hsiu Lin, Yu-Chen Hu, Chih-Hsien Hsia, Yi-An Lian, Sin-Ye Jhong
Summary: In smart cities, video surveillance based on a centralized framework may face limitations due to network connectivity in the cloud, hence the need for edge computing for real-time processing of media data.
Article
Computer Science, Theory & Methods
Yousung Yang, Seongsoo Lee, Joohyung Lee
Summary: This paper presents the design and implementation of a video analytics based real-time intelligent crossing detection system (RICDS) for smart cities. The system utilizes an adaptive queue management-based object tracking scheme to enhance object tracking on edge devices with limited computational resources. The system also introduces a real-world-real-time tracking scheme to predict the future positions of multiple objects and assign unique IDs to them. Experimental results show that the proposed tracking scheme achieves a significant latency reduction while maintaining similar multi object tracking accuracy compared to benchmark schemes.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2022)
Article
Computer Science, Information Systems
Ya-Jie Wu, Ricardo Brito, Wai-Hei Choi, Chi-Seng Lam, Man-Chung Wong, Sai-Weng Sin, Rui Paulo Martins
Summary: Smart meter monitors electricity consumption through modern metering devices connected to the IoT, providing intelligent and fast applications like arc fault protection based on nonintrusive monitoring load classification with fast safety responses. Traditional IoT architecture cannot support such fast responses under loading variation.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Automation & Control Systems
Jefferson Silva Almeida, Chenxi Huang, Fabricio Gonzalez Nogueira, Surbhi Bhatia, Victor Hugo C. de Albuquerque
Summary: This article proposes a novel lightweight convolutional neural network (CNN) model for wildfire detection through RGB images. The proposed method shows advantages in efficiency and accuracy, and can be combined with unmanned aerial vehicles and video surveillance systems for image processing. The ability to send timely wildfire alerts makes this method significant for forest protection.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Computer Science, Information Systems
Sudan Jha, Changho Seo, Eunmok Yang, Gyanendra Prasad Joshi
Summary: This paper introduces a system that enables real-time video surveillance in low-end edge computing environments by combining object detection tracking algorithm. The study proposes N-YOLO, a method that divides images into fixed-size sub-images and merges detection results using correlation-based tracking algorithm to significantly reduce computation for object detection and tracking. Additionally, a system is proposed to guarantee real-time performance in various edge computing environments by adaptively controlling the cycle of object detection and tracking.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Information Systems
S. Shitharth, Hariprasath Manoharan, Rakan A. Alsowail, Achyut Shankar, Saravanan Pandiaraj, Carsten Maple, Gwanggil Jeon
Summary: This article introduces the application of edge computing method and Internet of Things (IoT), which eliminates the need for cloud platform data processing by using offline IoT and incremental learning techniques. The proposed method combines edge computing, IoT, and incremental learning techniques to detect objects with varying weights.
INTERNET OF THINGS
(2023)
Article
Computer Science, Information Systems
Siyan Guo, Cong Zhao, Guiqin Wang, Jiaqing Yang, Shusen Yang
Summary: EC(2)Detect is a real-time online video object detection method that utilizes edge-cloud collaboration for accurate object detection by the Cloud and object tracking at edge devices. It significantly outperforms state-of-the-art methods in terms of processing speed, E2E latency, and edge-cloud bandwidth occupation, showing great potential for large-scale intelligent video analytics applications.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Chemistry, Multidisciplinary
Zuo Xiang, Patrick Seeling, Frank H. P. Fitzek
Summary: The separation of previously integrated inference from object recognition and other machine learning approaches has broad applicability in various scenarios. By splitting the inference component of the YOLOv2 trained machine learning model between client, network, and service, service latency can be reduced significantly while speed is increased.
APPLIED SCIENCES-BASEL
(2021)
Article
Computer Science, Information Systems
Yi Li, Zhangbing Zhou, Xiao Xue, Deng Zhao, Patrick C. K. Hung
Summary: This article proposes an accurate anomaly detection mechanism with energy efficiency in three-tier IoT-edge-cloud collaborative networks. It filters anomaly-relevant sensory data at the edge tier to decrease network traffic. The boundary of anomaly is determined using the Kriging spatial interpolation algorithm at the cloud tier and refined using mobile sensing nodes at edge networks.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Wuqin Tang, Qiang Yang, Xiaochen Hu, Wenjun Yan
Summary: This article proposes the use of unmanned aerial vehicles (UAVs) equipped with sensors and a cloud-based computer to detect defects in photovoltaic plants. The system utilizes an Internet of Things-based cloud-edge computing infrastructure to achieve low latency, low cost, and high accuracy defect detection. A two-stage algorithm is proposed to identify defects based on the characteristics of electroluminescence images.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Review
Computer Science, Hardware & Architecture
Yan Dai, Ziyu Hu, Shuqi Zhang, Lianjun Liu
Summary: This study comprehensively researches the problem of detection-based video multi-object tracking, studying the important aspects of MOT system, reviewing the classification and latest progress in object detection, and proposing possible research directions for the future.
Article
Automation & Control Systems
Fath U. Min Ullah, Khan Muhammad, Ijaz Ul Haq, Noman Khan, Ali Asghar Heidari, Sung Wook Baik, Victor Hugo C. de Albuquerque
Summary: This article proposes an AI-enabled IIoT-based framework called VD-Net to intelligently detect and analyze violent scenes, improving the accuracy of surveillance systems.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Marek Pawlicki, Aleksandra Pawlicka, Rafal Kozik, Michal Choras
Summary: Cloud computing, edge computing, and Internet-of-Things have significantly impacted people's lives, but their security should not be taken for granted. These paradigms are constantly under attack, and the potential breaches can have severe consequences. This systematic review aims to analyze the overlap of attacks in cloud, edge, and IoT and provide solutions and countermeasures to enhance their security. It fills the gap by constructing a concise threat catalogue and offering a more universal approach to ensure the safety of the entire ecosystem.
Review
Chemistry, Analytical
Mostafa Ahmed Ezzat, Mohamed A. Abd El Ghany, Sultan Almotairi, Mohammed A. -M. Salem
Summary: The automation strategy of today's smart cities heavily relies on large IoT systems for collecting big data analytics, particularly in the areas of video surveillance applications, algorithms, datasets, and embedded systems. This paper discusses the latest advancements in datasets, algorithms, and embedded systems for edge vision computing, while also addressing future trends and challenges in the field.
Article
Engineering, Electrical & Electronic
Hang Zou, Chao Zhang, Samson Lasaulce, Lucas Saludjian, H. Vincent Poor
Summary: This paper discusses the problem of executing a task from a quantized version of the information source. The task is modeled by minimizing a general goal function with quantized parameters. The paper shows how to design a quantizer to minimize the gap between the quantized version and the optimal result. The analysis provides quantization strategies and allows a practical algorithm to be designed.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2023)
Article
Automation & Control Systems
Xiaokang Zhou, Yiyong Hu, Jiayi Wu, Wei Liang, Jianhua Ma, Qun Jin
Summary: In this article, we propose a distribution bias aware collaborative GAN (DB-CGAN) model for imbalanced deep learning in industrial IoT. By introducing a complementary classifier and a data augmentation algorithm, the model can effectively handle the distribution bias between the generated data and the original data, resulting in significantly improved classification accuracy.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Automation & Control Systems
Ke Yan, Xinke Chen, Xiaokang Zhou, Zheng Yan, Jianhua Ma
Summary: Physics theory integrated machine learning models enhance the interpretability and performance of AI techniques for real-world industrial applications, such as AHU FDD. Traditional machine learning-based FDD models have high classification accuracy with sufficient training data but lack physical interpretation. This article presents a physical model integrated WGAN model for AHU FDD with insufficient training data. The proposed solution significantly improves model interpretability and outperforms existing methods.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Computer Science, Theory & Methods
Jiajun Li, Pu Wang, Long Jiao, Zheng Yan, Kai Zeng, Yishan Yang
Summary: Ambient backscatter communication (AmBC) is a cutting-edge technology that enables ultra-low-power communications for Internet of Things (IoT) applications by backscattering ambient radio frequency (RF) signals and harvesting energy simultaneously. However, existing research lacks effective secret key sharing schemes for safeguarding communications between resource-constrained backscatter devices (BDs) in AmBC systems. In this paper, a novel physical layer key generation scheme called Tri-Channel is proposed, which outperforms traditional schemes in terms of security and efficiency of key generation under passive and active attacks.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2023)
Article
Engineering, Electrical & Electronic
Ju-Hyung Lee, Hyowoon Seo, Jihong Park, Mehdi Bennis, Young-Chai Ko
Summary: This paper proposes a novel contention-based random access solution for low Earth orbit satellite (LEO SAT) networks, called eRACH, which achieves automatic protocol establishment through multi-agent deep reinforcement learning in a non-stationary network environment. In contrast to existing model-based and standardized protocols, eRACH does not require central coordination or additional communication across users, and training convergence is stabilized through regular orbiting patterns. Compared to RACH, simulation results show that eRACH achieves 54.6% higher average network throughput, around two times lower average access delay, and a Jain's fairness index of 0.989.
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2023)
Article
Computer Science, Hardware & Architecture
Won Joon Yun, Yunseok Kwak, Hankyul Baek, Soyi Jung, Mingyue Ji, Mehdi Bennis, Jihong Park, Joongheon Kim
Summary: This paper proposes a novel learning framework by integrating Federated Learning (FL) with width-adjustable slimmable neural networks (SNNs) to address the challenges posed by heterogeneous energy, wireless channel conditions, and non-IID data distributions. The proposed method, named SlimFL, utilizes superposition coding (SC) and superposition training (ST) to achieve communication and energy efficiency in global model aggregation and local model updating. Formal proofs and data-intensive simulations demonstrate that SlimFL is capable of dealing with non-IID data distributions and poor channel conditions while maintaining high communication efficiency.
IEEE-ACM TRANSACTIONS ON NETWORKING
(2023)
Article
Entomology
Sheng-Yu Zhang, Han Gao, Ankarjan Askar, Xing-Peng Li, Guo-Cai Zhang, Tian-Zhong Jing, Hang Zou, Hao Guan, Yun-He Zhao, Chuan-Shan Zou
Summary: This study reveals that the steroid hormone 20-hydroxyecdysone (20E) disrupts lipid metabolism in the fat body of Hyphantria cunea larvae, accelerating fatty acid beta-oxidation and promoting lipolysis. However, it negatively regulates gluconeogenesis.
Article
Automation & Control Systems
Xiaokang Zhou, Hailiang Hou, Wei Liang, Kevin I-Kai Wang, Qun Jin
Summary: In this paper, a double constrained containment mechanism is proposed for smart manufacturing to dispatch heterogeneous robots. A three-layer control framework is used to realize cloud-based collaborative manufacturing, which is more cost-effective and can handle dynamically changing communication topologies and delays. Theoretical stability analysis and experimental evaluation demonstrate the practicality of the proposed method in considering position and velocity constraints, switching topologies, and communication delays.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Computer Science, Hardware & Architecture
Xiaokang Zhou, Wei Liang, Kevin I-Kai Wang, Zheng Yan, Laurence T. Yang, Wei Wei, Jianhua Ma, Qun Jin
Summary: In this article, a Peer-to-Peer (P2P) based Privacy-Perceiving Asynchronous Federated Learning (PPAFL) framework is introduced for secure and resilient decentralized model training in modern mobile robotic systems. This framework uses reputation-aware coordination and secret sharing based communication mechanisms to achieve encrypted P2P federated learning and anonymous local model updates.
IEEE WIRELESS COMMUNICATIONS
(2023)
Article
Computer Science, Information Systems
Zhonghong Ou, Zhaofengnian Wang, Fenrui Xiao, Baiqiao Xiong, Hongxing Zhang, Meina Song, Yan Zheng, Pan Hui
Summary: With the popularity of 5G networks and IoT applications, real-time environmental awareness becomes crucial. However, small object detection still faces challenges due to limited scales and low detection accuracy. To address these issues, the proposed AD-RCNN employs dynamic region proposal network, visual attention scheme, and adaptive dynamic training module. Experimental results demonstrate that AD-RCNN outperforms existing methods in terms of mAP and FPS.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Cardiac & Cardiovascular Systems
Xuyuan Kuang, Zihao Zhong, Wei Liang, Suzhen Huang, Renji Luo, Hui Luo, Yongheng Li
Summary: This paper analyzes the application of machine learning in heart failure-associated diseases using bibliometric methods and provides a comprehensive overview of heart failure-related machine learning publications. The study screened articles from Web of Science and employed intuitive data analysis and VOSViewer for analysis. The top 100 most cited articles were also investigated.
FRONTIERS IN CARDIOVASCULAR MEDICINE
(2023)
Article
Engineering, Civil
Yan Chen, Tian Shu, Xiaokang Zhou, Xuzhe Zheng, Akira Kawai, Kaoru Fueda, Zheng Yan, Wei Liang, Kevin I-Kai Wang
Summary: In this paper, a Graph Attention Network with Spatial-Temporal Clustering (GAT-STC) is proposed to improve traffic flow forecasting in Intelligent Transportation System (ITS). The network extracts recent-aware features and periodic-aware features to capture dynamic changes in spatial feature representation. Experimental results show that the proposed model outperforms five baseline methods in terms of accuracy and efficiency.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Automation & Control Systems
Xinmeng Liu, Haomeng Xie, Zheng Yan, Xueqin Liang
Summary: This paper proposes evaluation criteria regarding scalability, applicability, and reliability in blockchain. It also classifies and provides an overview of advanced sharding techniques, analyzing their respective advantages and disadvantages. The paper highlights unresolved issues and suggests future research directions.
Article
Computer Science, Information Systems
David E. E. Ruiz-Guirola, Onel L. A. Lopez, Samuel Montejo-Sanchez, Richard Demo Souza, Mehdi Bennis
Summary: Prolonging the lifetime of MTC networks is crucial for a sustainable digitized society. Accurately predicting MTC traffic and optimizing resource allocation can lead to significant energy savings. However, selecting the right predictor depends on trade-offs between accuracy, complexity, and network characteristics, and this debate is lacking in current literature.
IEEE WIRELESS COMMUNICATIONS LETTERS
(2023)
Proceedings Paper
Computer Science, Hardware & Architecture
Mounssif Krouka, Anis Elgabli, Chaouki ben Issaid, Mehdi Bennis
Summary: In this paper, a decentralized Newton-type approach is proposed to solve the problem of decentralized federated learning. The algorithm leverages the fast convergence of second-order methods and reduces communication and privacy concerns. The approach consists of solving an inner problem and an outer problem alternately using a decentralized manner and performing one decentralized Newton step at every iteration. Simulation results demonstrate that the proposed algorithm outperforms several baselines and provides efficient solutions for bandwidth-limited systems under different SNR regimes.
2023 IEEE WIRELESS COMMUNICATIONS AND NETWORKING CONFERENCE, WCNC
(2023)