Article
Computer Science, Information Systems
Dewang Ren, Xiaolin Gui, Kaiyuan Zhang
Summary: This article proposes a twin-timescale framework to jointly optimize adaptive request scheduling and cooperative service caching in multidevice and MEC-assisted networks, addressing various challenges. By modeling the request scheduling and service caching problems as POMDP problems and using an online algorithm based on deep reinforcement learning, it effectively improves service latency and hit rate, further reducing average service latency.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Zhigang Xie, Xin Song, Jing Cao, Weipeng Qiu
Summary: The evolution of information and communication technology has led to the emergence of the Internet of Things and the challenges of limited battery power and lack of infrastructure in emergency scenarios. Tethered unmanned aerial vehicles (UAVs) are proposed as an alternative base station to provide communication and task offloading services in areas without infrastructure.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Xuemei Yang, Hong Luo, Yan Sun, Mohsen Guizani
Summary: This paper studies the optimization problem of task offloading in multiaccess edge computing (MEC), and proposes a hybrid average reward proximal policy optimization (hybrid-ARPPO) algorithm to jointly optimize offloading decisions, cooperative transmission ratios, and edge server assignments, aiming to reduce the long-term average task execution cost.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Jaeyoung Hwang, Lionel Nkenyereye, Nakmyoung Sung, Jaeho Kim, Jaeseung Song
Summary: As IoT technology advances, various industries benefit from its value-added services, however, the simple integration of advanced technologies may not fully utilize their advantages. Therefore, there is a need for an efficient integration mechanism. The new architectural framework virtualizes an IoT platform at the edge node to support specific IoT services, reducing latency and efficiently managing IoT services.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Computer Science, Information Systems
Anahita Mazloomi, Hani Sami, Jamal Bentahar, Hadi Otrok, Azzam Mourad
Summary: This article addresses the problem of minimizing network delay and the number of edge servers in multiaccess edge computing (MEC) design. It proposes a novel reinforcement learning framework and demonstrates its effectiveness through experiments with real-world datasets.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Chemistry, Analytical
Adrian Orive, Aitor Agirre, Hong-Linh Truong, Isabel Sarachaga, Marga Marcos
Summary: The fast growth in the amount of connected devices has enabled the emergence of Edge computing, which provides lower latencies compared to Cloud computing. Combining Cloud and Edge computing can meet the quality of service requirements for complex applications. However, orchestrating applications in the Cloud-Edge computing faces new challenges that need to be solved to fully utilize this layered infrastructure. This paper proposes an architecture that dynamically orchestrates applications in the Cloud-Edge continuum, focusing on the application's quality of service.
Article
Computer Science, Information Systems
Giovanni Nardini, Alessandro Noferi, Giovanni Stea
Summary: Multi-access Edge Computing (MEC) is proposed to enable platooning of vehicles, utilizing mobile networks for connectivity. However, as platoons consist of vehicles from different mobile operators, customers of different MEC systems, this paper presents an architectural framework called Platooning-as-a-Service (PlaaS) in a multi-operator MEC environment, compliant with the ETSI MEC standard. The framework is evaluated using an open-source proof-of-concept implementation in the Simu5G simulator, assessing its impact on platoon stability, particularly the latencies.
Article
Computer Science, Information Systems
Xianbang Diao, Xinrong Guan, Yueming Cai
Summary: This study utilizes multiaccess edge computing (MEC) to assist Internet of Things (IoT) devices in completing complex status updates. It proposes a joint optimization problem that aims to minimize average peak Age of Information (AoI), average energy consumption of IoT devices, and average energy consumption of UAVs. The results show that the proposed algorithm effectively improves system performance by achieving the desired tradeoff among the three performance metrics.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Hardware & Architecture
Claudio Correia, Miguel Correia, Luis Rodrigues
Summary: This article presents the design and implementation of a secure event ordering service for fog nodes. The service leverages a Trusted Execution Environment (TEE) to provide guarantees regarding the order of events, even when fog nodes are compromised.
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING
(2022)
Article
Computer Science, Information Systems
Zhen Qin, Hai Wang, Zhenhua Wei, Yuben Qu, Fei Xiong, Haipeng Dai, Tao Wu
Summary: The article focuses on task selection and scheduling in UAV-enabled multiaccess edge computing for reconnaissance. It proposes a solution to maximize reconnaissance utility by optimizing task selection and scheduling sequence, addressing a challenging mixed-integer nonlinear programming problem. The proposed algorithm improves overall reconnaissance utility and energy efficiency compared to benchmark algorithms.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Computer Science, Information Systems
Ying Li, Xingwei Wang, Rongfei Zeng, Mingzhou Yang, Kexin Li, Min Huang, Schahram Dustdar
Summary: This article proposes an incentive mechanism named VARF based on the infinitely repeated game for cross-silo federated learning. VARF aims to address the challenge of high-quality client selection and long-term client participation in model training for efficient and stable FL in multiaccess edge computing (MEC). Experimental results demonstrate the superiority of VARF compared to other benchmarks, and it also shows that a long and stable cooperative relationship can be formed between cloud platforms and edge nodes.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Kexin Li, Xingwei Wang, Qiang He, Qiang Ni, Mingzhou Yang, Schahram Dustdar
Summary: This article introduces a new computational offloading decision framework that minimizes the long-term payment of computational tasks with mixed bound constraints. It addresses the issue of tasks exceeding soft constraints and improves learning efficiency and convergence speed through the elimination of redundant experiences and optimal transitions.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Minseok Song, Yeongju Lee, Kyungmin Kim
Summary: The study proposes a scheme in multiaccess edge computing where tasks are offloaded from mobile devices to servers, aiming to maximize reward within limited power budget, server processing capacities, and wireless network coverage.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Computer Science, Information Systems
Shu Yang, Kunkun Xu, Laizhong Cui, Zhongxing Ming, Ziteng Chen, Zhong Ming
Summary: Edge computing, especially multiaccess edge computing, is considered a promising technology for improving user experience in many AI applications towards IoT infrastructure. The proposed EBI-PAI platform, based on SDN and serverless technology, provides a unified service calling interface and automatic resource scheduling to meet user QoE requirements. Comprehensive simulations and a case study show that EBI-PAI can greatly enhance QoE while saving budget, validating the effectiveness of the platform.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Computer Science, Hardware & Architecture
Mohammed Laroui, Hatem Ibn Khedher, Hassine Moungla, Hossam Afifi
Summary: This paper presents new approaches for Service Function Chain (SFC) orchestration in Virtual Mobile Edge Computing (VMEC) environments in IoT networks. The proposed Optimal SFC Placement in VMEC over AI-IoT (OSPV) algorithm and Efficient SFC Placement in VMEC over AI-IoT (ESPV) algorithm address the balance between resource optimization and performance improvement. Deep Learning techniques are used to predict mobility and energy consumption sequences, which enable efficient SFC placement. Experimental results show the effectiveness of the proposed algorithms in different computing scenarios.
Article
Computer Science, Artificial Intelligence
Heena Rathore, Amr Mohamed, Mohsen Guizani, Shailendra Rathore
Summary: This paper introduces a machine learning approach called NueroFATH for the physical assessment of athletes. It uses neural networks and fuzzy c-means techniques to predict the potential of athletes winning medals. The study also identifies important physical characteristics related to the assessment results.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Ali Riahi, Amr Mohamed, Aiman Erbad
Summary: A decentralized Blockchain-based framework is proposed to guarantee users' mutual trust while sharing their local learning experiences. The Reinforcement Learning-based Federated Learning framework is optimized to improve global learning performance. The efficacy of the proposed framework is demonstrated in a realistic containerized environment.
IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT
(2023)
Article
Computer Science, Information Systems
Muhammad Asif Khan, Ridha Hamila, Aiman Erbad, Moncef Gabbouj
Summary: This article introduces a method for distributed inference over IoT devices and edge server, reducing computational load and energy consumption by employing early split strategy and late split strategy.
IEEE SYSTEMS JOURNAL
(2023)
Article
Computer Science, Information Systems
Mohammed Seid Abegaz, Hayla Nahom Abishu, Yasin Habtamu Yacob, Tewodros Alemu Ayall, Aiman Erbad, Mohsen Guizani
Summary: The use of drones as aerial base stations in the Industrial Internet of Things (IIoT) network has been proposed to address the increasing demands for low-latency data communication from mobile devices. A novel intelligent resource trading framework is proposed, which utilizes multi-agent deep reinforcement learning, blockchain, and game theory to manage dynamic resource trading environments. Simulation results demonstrate the effectiveness of the proposed scheme in improving resource trading in UAV-assisted IIoT networks.
IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT
(2023)
Review
Social Sciences, Interdisciplinary
Yassine Himeur, Somaya Al-Maadeed, Iraklis Varlamis, Noor Al-Maadeed, Khalid Abualsaud, Amr Mohamed
Summary: This paper reviews the progress in face mask detection research, with a focus on deep learning and deep transfer learning techniques. It first describes and discusses existing face mask detection datasets, then presents recent advances in object detectors and Convolutional Neural Network architectures, as well as the different deep learning techniques that have been applied. Benchmarking results are summarized, and the limitations of datasets and methodologies are discussed. Lastly, future research directions are discussed in detail.
Article
Automation & Control Systems
Abdulrahman Soliman, Abdulla Al-Ali, Amr Mohamed, Hend Gedawy, Daniel Izham, Mohamad Bahri, Aiman Erbad, Mohsen Guizani
Summary: In this paper, a novel framework for UAV smart navigation is proposed to minimize time and energy consumption. Deep Reinforcement Learning is used to enable the drone to learn target mobility patterns and develop energy-efficient scanning strategies. Simulation and hardware integration testbed validate the feasibility and practicality of the proposed methodology.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Muhammad Asif Khan, Emna Baccour, Aiman Erbad, Ridha Hamila, Mounir Hamdi
Summary: Video traffic over the Internet is rapidly increasing, projected to account for 82% of total Internet traffic by 2022. The existing network architectures face challenges due to this significant growth, especially with the inclusion of immersive video content. Mobile edge computing (MEC) is seen as a solution to improving the user experience. However, it is important to address the limitation of edge servers in handling sudden spikes in video traffic. This article proposes a novel D2D-MEC collaborative framework called CODE (Computation Offloading in D2D-Edge) that offloads computations to distributed user devices, offering two schemes and a heuristic to solve the computation offloading problem and improve edge resource utilization.
IEEE SYSTEMS JOURNAL
(2023)
Article
Engineering, Electrical & Electronic
Maregu Assefa, Wei Jiang, Kumie Gedamu Alemu, Getinet Yilma, Deepak Adhikari, Melese Ayalew, Abegaz Mohammed Seid, Aiman Erbad
Summary: Self-supervised contrastive learning has made significant progress in action recognition tasks, but existing video datasets have biased the learned representations towards dominant backgrounds and scene correlations. Our proposed Actor-aware Self-supervised Learning (ActorSL) addresses this issue by aligning actors and scenes, introducing background mixing augmentation, and optimizing contrastive loss and consistency regularization in a semi-supervised manner. Experimental results show that ActorSL outperforms current state-of-the-art semi-supervised methods on multiple datasets.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Automation & Control Systems
Abdullatif Albaseer, Mohamed Abdallah, Ala Al-Fuqaha, Aiman Erbad
Summary: Federated edge learning (FEEL) is a distributed learning technique for next-generation wireless edge systems. This article presents joint participant selection and bandwidth allocation schemes to address challenges including data heterogeneity and limited wireless resources. Experimental results demonstrate that our proposed scheme improves convergence rate by up to 55% compared to benchmarks.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Computer Science, Information Systems
Sergio Marquez-Sanchez, Jaime Calvo-Gallego, Aiman Erbad, Muhammad Ibrar, Javier Hernandez Fernandez, Mahdi Houchati, Juan Manuel Corchado
Summary: This article introduces a cutting-edge edge computing architecture based on virtual organizations, federated learning, and deep reinforcement learning algorithms, aiming to optimize energy consumption in buildings and homes and address the cybersecurity risks and data transmission overheads associated with cloud-based systems.
Article
Telecommunications
Abegaz Mohammed Seid, Aiman Erbad, Hayla Nahom Abishu, Abdullatif Albaseer, Mohamed Abdallah, Mohsen Guizani
Summary: This paper proposes a dynamic resource allocation framework that synergies blockchain and multi-agent deep reinforcement learning for multi-UAV-enabled 5G-RAN to allocate resources to smart mobile user equipment (SMUE) with optimal costs. The blockchain ensures the security of virtual resource transactions between SMUEs, infrastructure providers (InPs), and virtual network operators (VNOs). The simulation results show that the proposed method outperforms the state-of-the-art methods in terms of utility optimization and quality of service satisfaction.
IEEE TRANSACTIONS ON COGNITIVE COMMUNICATIONS AND NETWORKING
(2023)
Review
Engineering, Civil
Afshan Ahmed, Muhammad Munwar Iqbal, Sohail Jabbar, Muhammad Ibrar, Aiman Erbad, Houbing Song
Summary: In recent years, significant interest has been shown in vehicular networks to improve road safety through real-time messaging services among vehicles. This work provides a detailed analysis of emergency message dissemination techniques for the Internet of Vehicles (IoV). Position-based data dissemination techniques are explored, which are considered the best routing method due to their independence from predestination entries. This article differs from existing survey papers by examining a brief comparison of beacon-oriented and beacon-less techniques for position-based emergency message routing.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Mhd Saria Allahham, Amr Mohamed, Aiman Erbad, Mohsen Guizani
Summary: Mobile edge learning (MEL) is a learning paradigm that enables distributed training of machine learning models over heterogeneous edge devices. This study proposes an incentive mechanism to motivate the participation of edge devices in the training process and evaluates its performance through numerical experiments.
IEEE CANADIAN JOURNAL OF ELECTRICAL AND COMPUTER ENGINEERING
(2023)
