Hailong Zhu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Hailong Zhu has a total of 62 co-authors. Here are the top co-authors by number of shared publications:
- Peiying Zhang
China University of Petroleum (East China)
11 shared publications - Ning Chen
School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China
3 shared publications - Bin Liang
Tsinghua University
2 shared publications - Tao Huang
Beijing University of Posts and Telecommunications
2 shared publications - Tao Zhang
Tsinghua University
2 shared publications - Yi Zhang
Qingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China), Qingdao, China
2 shared publications
Latest Publications
- 7 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (5) , 4570-4578
Explainable Edge AI Framework for IoD-Assisted Aerial Surveillance in Extreme Scenarios
Drones are sophisticated machines that can hover over extreme locations, conduct aerial surveillance, collect surveillance data, and disseminate it to the distributed edge for processing and analysis. The distributed edge deploys advanced Artificial Intelligence (AI) models to detect any unwarranted activity or object based on surveillance data. However, these lightweight and low-power Unmanned Aerial Vehicles (UAVs) may experience faults due to unprecedented workload when deployed in extreme surveillance domains. In this paper, we have designed an AI framework to detect any safety concerns with drones deployed for aerial surveillance in extreme locations based on real-time drone critical parameters. We also propose a MapReduce-based object recognition and classification module to process large-scale images captured by drones efficiently. However, conventional AI systems behave like black box systems, leading to a lack of trust and transparency. Thus, we convert the traditional framework of AI into an explainable edge AI framework using SHapley Additive exPlanations (SHAP) that opens Pandora’s black box. The experimental results show the effectiveness of the proposed framework in detecting drone safety concerns through explainable health status tracking alongside ensuring an effective object detection mechanism.
- 7 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12 (5) , 4570-4578
Explainable Edge AI Framework for IoD-Assisted Aerial Surveillance in Extreme Scenarios
Drones are sophisticated machines that can hover over extreme locations, conduct aerial surveillance, collect surveillance data, and disseminate it to the distributed edge for processing and analysis. The distributed edge deploys advanced Artificial Intelligence (AI) models to detect any unwarranted activity or object based on surveillance data. However, these lightweight and low-power Unmanned Aerial Vehicles (UAVs) may experience faults due to unprecedented workload when deployed in extreme surveillance domains. In this paper, we have designed an AI framework to detect any safety concerns with drones deployed for aerial surveillance in extreme locations based on real-time drone critical parameters. We also propose a MapReduce-based object recognition and classification module to process large-scale images captured by drones efficiently. However, conventional AI systems behave like black box systems, leading to a lack of trust and transparency. Thus, we convert the traditional framework of AI into an explainable edge AI framework using SHapley Additive exPlanations (SHAP) that opens Pandora’s black box. The experimental results show the effectiveness of the proposed framework in detecting drone safety concerns through explainable health status tracking alongside ensuring an effective object detection mechanism.
- 24 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (4) , 6814-6824
Virtual Network Embedding for Task Offloading in IIoT: A DRL-Assisted Federated Learning Scheme
The Industrial Internet of Things (IIoT) promotes the deep integration of new-generation communication technologies and industrial ecology. However, the popularity of computing and the proliferation of equipment scale make it a meaningful challenge to provide reasonable resource allocation for task offloading. Therefore, this article proposes a novel two-stage coordinated, distributed, and online multidomain virtual network embedding algorithm based on deep reinforcement learning (DRL)-assisted federated learning (FL) for task offloading in the IIoT. We model the IIoT as a dynamic multidomain structure and deploy local DRL servers in each factory domain combined with the distributed paradigm of FL to reduce the local resource fragmentation. Through local and global cooperation, the IIoT environment is controlled in a fine and macroscopic manner. In addition, the mechanisms of FL ensure the privacy of participant data. Finally, a comprehensive evaluation demonstrates the clear superiority of the proposed algorithm, which improves the long-term offloading revenue, resource utilization, and task offloading success rate by average 17.66%, 5.97%, and 4.52% compared to baselines, respectively.
- 24 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (4) , 6814-6824
Virtual Network Embedding for Task Offloading in IIoT: A DRL-Assisted Federated Learning Scheme
The Industrial Internet of Things (IIoT) promotes the deep integration of new-generation communication technologies and industrial ecology. However, the popularity of computing and the proliferation of equipment scale make it a meaningful challenge to provide reasonable resource allocation for task offloading. Therefore, this article proposes a novel two-stage coordinated, distributed, and online multidomain virtual network embedding algorithm based on deep reinforcement learning (DRL)-assisted federated learning (FL) for task offloading in the IIoT. We model the IIoT as a dynamic multidomain structure and deploy local DRL servers in each factory domain combined with the distributed paradigm of FL to reduce the local resource fragmentation. Through local and global cooperation, the IIoT environment is controlled in a fine and macroscopic manner. In addition, the mechanisms of FL ensure the privacy of participant data. Finally, a comprehensive evaluation demonstrates the clear superiority of the proposed algorithm, which improves the long-term offloading revenue, resource utilization, and task offloading success rate by average 17.66%, 5.97%, and 4.52% compared to baselines, respectively.
- Open Access
- 30 September 2023
- Published by MDPI AG in Electronics
- Vol. 12 (19) , 4103
Flow-Based Joint Programming of Time Sensitive Task and Network
Owning to the application of artificial intelligence and big data analysis in industry, automobiles, aerospace, and other fields, the high-bandwidth candidate, time-sensitive networking (TSN), is introduced into the data communication network. Apart from keeping the safety-critical and real-time requirements, it faces challenges to satisfy large traffic transmission, such as sampled video for computer vision. In this paper, we consider task scheduling and time-sensitive network together and formalize them into a first-order-constraints satisfy module theory (SMT) problem. Based on the result of the solver, we build flow-level scheduling based on IEEE 802.1 Qbv. By splitting the flow properly, it can reduce the constraint inequality as the traffic grows more than the traditional frame-based programming model and achieve near 100% utilization. It can be a general model for the deterministic task and network scheduling design.
- 30 September 2023
- Published by MDPI AG in Electronics
- Vol. 12 (19) , 4103
Open AccessFlow-Based Joint Programming of Time Sensitive Task and Network
Owning to the application of artificial intelligence and big data analysis in industry, automobiles, aerospace, and other fields, the high-bandwidth candidate, time-sensitive networking (TSN), is introduced into the data communication network. Apart from keeping the safety-critical and real-time requirements, it faces challenges to satisfy large traffic transmission, such as sampled video for computer vision. In this paper, we consider task scheduling and time-sensitive network together and formalize them into a first-order-constraints satisfy module theory (SMT) problem. Based on the result of the solver, we build flow-level scheduling based on IEEE 802.1 Qbv. By splitting the flow properly, it can reduce the constraint inequality as the traffic grows more than the traditional frame-based programming model and achieve near 100% utilization. It can be a general model for the deterministic task and network scheduling design.
- Open Access
A Collaborative Inference Algorithm in Low-Earth-Orbit Satellite Network for Unmanned Aerial Vehicle
In recent years, the low-Earth-orbit (LEO) satellite network has achieved considerable development. Moreover, it is necessary to introduce edge computing into LEO networks, which can provide high-quality services, such as worldwide seamless low-delay computation offloading for unmanned aerial vehicles (UAVs) or user terminals and nearby remote-sensing data processing for UAVs or satellites. However, because the computation resource of the satellite is relatively scarce compared to the ground server, it is hard for a single satellite to complete massive deep neural network (DNN) inference tasks in a short time. Consequently, in this paper, we focus on the multi-satellite collaborative inference problem and propose a novel COllaborative INference algorithm for LEO edge computing called COIN-LEO. COIN-LEO manages to split the complete DNN model into several submodels consisting of some consecutive layers and deploy these submodels to several satellites for inference. We innovatively leverage deep reinforcement learning (DRL) to efficiently split the model and use a neural network (NN) to predict the time required for inference tasks of a specific submodel on a specific satellite. By implementing COIN-LEO and evaluating its performance in a highly realistic satellite-network-emulation platform, we find that our COIN-LEO outperforms baseline algorithms in terms of inference throughput, time consumed and network traffic overhead.
Open AccessA Collaborative Inference Algorithm in Low-Earth-Orbit Satellite Network for Unmanned Aerial Vehicle
In recent years, the low-Earth-orbit (LEO) satellite network has achieved considerable development. Moreover, it is necessary to introduce edge computing into LEO networks, which can provide high-quality services, such as worldwide seamless low-delay computation offloading for unmanned aerial vehicles (UAVs) or user terminals and nearby remote-sensing data processing for UAVs or satellites. However, because the computation resource of the satellite is relatively scarce compared to the ground server, it is hard for a single satellite to complete massive deep neural network (DNN) inference tasks in a short time. Consequently, in this paper, we focus on the multi-satellite collaborative inference problem and propose a novel COllaborative INference algorithm for LEO edge computing called COIN-LEO. COIN-LEO manages to split the complete DNN model into several submodels consisting of some consecutive layers and deploy these submodels to several satellites for inference. We innovatively leverage deep reinforcement learning (DRL) to efficiently split the model and use a neural network (NN) to predict the time required for inference tasks of a specific submodel on a specific satellite. By implementing COIN-LEO and evaluating its performance in a highly realistic satellite-network-emulation platform, we find that our COIN-LEO outperforms baseline algorithms in terms of inference throughput, time consumed and network traffic overhead.
- Open Access
- 3 August 2023
- Published by MDPI AG in Electronics
- Vol. 12 (15) , 3330
Network Resource Allocation Algorithm Using Reinforcement Learning Policy-Based Network in a Smart Grid Scenario
The exponential growth in user numbers has resulted in an overwhelming surge in data that the smart grid must process. To tackle this challenge, edge computing emerges as a vital solution. However, the current heuristic resource scheduling approaches often suffer from resource fragmentation and consequently get stuck in local optimum solutions. This paper introduces a novel network resource allocation method for multi-domain virtual networks with the support of edge computing. The approach entails modeling the edge network as a multi-domain virtual network model and formulating resource constraints specific to the edge computing network. Secondly, a policy network is constructed for reinforcement learning (RL) and an optimal resource allocation strategy is obtained under the premise of ensuring resource requirements. In the experimental section, our algorithm is compared with three other algorithms. The experimental results show that the algorithm has an average increase of 5.30%, 8.85%, 15.47% and 22.67% in long-term average revenue–cost ratio, virtual network request acceptance ratio, long-term average revenue and CPU resource utilization, respectively.
- 3 August 2023
- Published by MDPI AG in Electronics
- Vol. 12 (15) , 3330
Open AccessNetwork Resource Allocation Algorithm Using Reinforcement Learning Policy-Based Network in a Smart Grid Scenario
The exponential growth in user numbers has resulted in an overwhelming surge in data that the smart grid must process. To tackle this challenge, edge computing emerges as a vital solution. However, the current heuristic resource scheduling approaches often suffer from resource fragmentation and consequently get stuck in local optimum solutions. This paper introduces a novel network resource allocation method for multi-domain virtual networks with the support of edge computing. The approach entails modeling the edge network as a multi-domain virtual network model and formulating resource constraints specific to the edge computing network. Secondly, a policy network is constructed for reinforcement learning (RL) and an optimal resource allocation strategy is obtained under the premise of ensuring resource requirements. In the experimental section, our algorithm is compared with three other algorithms. The experimental results show that the algorithm has an average increase of 5.30%, 8.85%, 15.47% and 22.67% in long-term average revenue–cost ratio, virtual network request acceptance ratio, long-term average revenue and CPU resource utilization, respectively.
- 1 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Broadcasting
- Vol. 69 (4) , 872-881
Machine Learning Assisted Video Stream Offloading for 5G MBMS Mobile Edge Computing
With the advancement of multimedia technology, 5G MBMS has become a promising technology that makes full and effective use of network resources to provide services. Mobile edge computing (MEC) is a key technology that brings important improvements to MBMS by utilizing contextual information such as multimedia, virtual reality, etc. However, the deployment of MEC may bring the latency to the network, which requires the deployment consideration of computing nodes. Moreover, mobile network operators need to consider the number of base stations required to accommodate the upcoming MBMS data streams. In this paper, we propose a computational offloading strategy and a co-optimization mechanism of video resource assignment in an edge cloud server enabled MBMS service system. Our approach aims to maximize computation speed while satisfying limited resource constraints, constrained energy consumption, and minimum latency time requirements, in order to provide better MBMS services. Additionally, our approach offers an advanced solution for achieving optimal computational offloading and resource allocation in edge cloud server enabled MBMS service systems. Extensive simulation results demonstrate that the proposed mechanism outperforms current state-of-the-art schemes. Specifically, the total calculation rate of the large-scale cloud server is significantly reduced under energy constraints, total delay requirements, and different network parameters.
- 1 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Broadcasting
- Vol. 69 (4) , 872-881
Machine Learning Assisted Video Stream Offloading for 5G MBMS Mobile Edge Computing
With the advancement of multimedia technology, 5G MBMS has become a promising technology that makes full and effective use of network resources to provide services. Mobile edge computing (MEC) is a key technology that brings important improvements to MBMS by utilizing contextual information such as multimedia, virtual reality, etc. However, the deployment of MEC may bring the latency to the network, which requires the deployment consideration of computing nodes. Moreover, mobile network operators need to consider the number of base stations required to accommodate the upcoming MBMS data streams. In this paper, we propose a computational offloading strategy and a co-optimization mechanism of video resource assignment in an edge cloud server enabled MBMS service system. Our approach aims to maximize computation speed while satisfying limited resource constraints, constrained energy consumption, and minimum latency time requirements, in order to provide better MBMS services. Additionally, our approach offers an advanced solution for achieving optimal computational offloading and resource allocation in edge cloud server enabled MBMS service systems. Extensive simulation results demonstrate that the proposed mechanism outperforms current state-of-the-art schemes. Specifically, the total calculation rate of the large-scale cloud server is significantly reduced under energy constraints, total delay requirements, and different network parameters.
- Open Access
UAV Ad Hoc Network Routing Algorithms in Space–Air–Ground Integrated Networks: Challenges and Directions
With the rapid development of 5G and 6G communications in recent years, there has been significant interest in space–air–ground integrated networks (SAGINs), which aim to achieve seamless all-area, all-time coverage. As a key component of SAGINs, flying ad hoc networks (FANETs) have been widely used in the agriculture and transportation sectors in recent years. Reliable communication in SAGINs requires efficient routing algorithms to support them. In this study, we analyze the unique communication architecture of FANETs in SAGINs. At the same time, existing routing protocols are presented and clustered. In addition, we review the latest research advances in routing algorithms over the last five years. Finally, we clarify the future research trends of FANET routing algorithms in SAGINs by discussing the algorithms and comparing the routing experiments with the characteristics of unmanned aerial vehicles.
Open AccessUAV Ad Hoc Network Routing Algorithms in Space–Air–Ground Integrated Networks: Challenges and Directions
With the rapid development of 5G and 6G communications in recent years, there has been significant interest in space–air–ground integrated networks (SAGINs), which aim to achieve seamless all-area, all-time coverage. As a key component of SAGINs, flying ad hoc networks (FANETs) have been widely used in the agriculture and transportation sectors in recent years. Reliable communication in SAGINs requires efficient routing algorithms to support them. In this study, we analyze the unique communication architecture of FANETs in SAGINs. At the same time, existing routing protocols are presented and clustered. In addition, we review the latest research advances in routing algorithms over the last five years. Finally, we clarify the future research trends of FANET routing algorithms in SAGINs by discussing the algorithms and comparing the routing experiments with the characteristics of unmanned aerial vehicles.
- 28 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cloud Computing
- Vol. 11 (4) , 3426-3438
Joint Trajectory and Energy Consumption Optimization Based on UAV Wireless Charging in Cloud Computing System
Microwave Power Transfer (MPT) is a promising technology to charge sensor devices (SDs) wirelessly in wireless sensor networks, and Cloud Computing (CC) can significantly promote task processing capacity of SDs. However, the propagation loss can dramatically influence the harvested energy and computation performance. So, for wireless sensor networks, we study an unmanned aerial vehicle-assisted cloud wireless charging system with the cooperation of the cloud server and the unmanned aerial vehicle (UAV). First, the UAV acts as the energy transmitter, and we design a quantitative charging scheme according to the energy-aware of SDs' battery capacity. Secondly, the cloud server processes the tasks uploaded by SDs with the cooperation of the UAV, and we consider the communication connection between the cloud server and the UAV. Third, we propose the joint resource-trajectory optimization to reduce the energy consumption of UAVs. We put forward the Chaotically Adaptive Beetle Swarm Optimization Based on Cauchy Mutation (CABSOC) assisted block coordinate descent algorithm for addressing this non-convex problem. Numerical results indicate that the proposed solution can significantly improve the energy performance of the UAV. And the energy consumption is reduced by 11% compared with the solution with network function virtualization (NFV).
- 28 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cloud Computing
- Vol. 11 (4) , 3426-3438
Joint Trajectory and Energy Consumption Optimization Based on UAV Wireless Charging in Cloud Computing System
Microwave Power Transfer (MPT) is a promising technology to charge sensor devices (SDs) wirelessly in wireless sensor networks, and Cloud Computing (CC) can significantly promote task processing capacity of SDs. However, the propagation loss can dramatically influence the harvested energy and computation performance. So, for wireless sensor networks, we study an unmanned aerial vehicle-assisted cloud wireless charging system with the cooperation of the cloud server and the unmanned aerial vehicle (UAV). First, the UAV acts as the energy transmitter, and we design a quantitative charging scheme according to the energy-aware of SDs' battery capacity. Secondly, the cloud server processes the tasks uploaded by SDs with the cooperation of the UAV, and we consider the communication connection between the cloud server and the UAV. Third, we propose the joint resource-trajectory optimization to reduce the energy consumption of UAVs. We put forward the Chaotically Adaptive Beetle Swarm Optimization Based on Cauchy Mutation (CABSOC) assisted block coordinate descent algorithm for addressing this non-convex problem. Numerical results indicate that the proposed solution can significantly improve the energy performance of the UAV. And the energy consumption is reduced by 11% compared with the solution with network function virtualization (NFV).
- 10 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 72 (8) , 10724-10736
A Proactive Stable Scheme for Vehicular Collaborative Edge Computing
Due to the restricted computing resources and high upgrading costs, onboard processors alone cannot meet the quality of service (QoS) requirements of the emerging and constantly upgrading vehicular applications. Computation offloading is a feasible solution to the excessive computation-intensive tasks. Meanwhile, vehicular collaborative edge computing (VCEC) is a paradigm to make the best use of surrounding vehicles' idle computing resources when the task density on a specific road segment suddenly increases. However, the high mobility of vehicles and the ad hoc nature of vehicular networks make maintaining stable task offloading performance quite challenging. Especially when the idle computing resources surrounding vehicles are insufficient, there has not been researching to achieve stable offloading performance. Based on this, we consider a proactive strategy that can decrease the events that affect performance stability. First, we propose a mobility prediction model for future network status prediction. Then we design an adaptive task offloading scheme based on proactive adjusting (PATO) to maintain stable task offloading performance. The scheme includes a state processing model and a deep reinforcement learning (DRL)-based task offloading algorithm. Finally, we conduct extensive simulations in various scenarios with insufficient resources to validate the task offloading performance of PATO. Compared with the existing DRL-based algorithm, the simulation results show that PATO can improve the mean offloading utility by 95.4% and the completion ratio by 15.8%.
- 10 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 72 (8) , 10724-10736
A Proactive Stable Scheme for Vehicular Collaborative Edge Computing
Due to the restricted computing resources and high upgrading costs, onboard processors alone cannot meet the quality of service (QoS) requirements of the emerging and constantly upgrading vehicular applications. Computation offloading is a feasible solution to the excessive computation-intensive tasks. Meanwhile, vehicular collaborative edge computing (VCEC) is a paradigm to make the best use of surrounding vehicles' idle computing resources when the task density on a specific road segment suddenly increases. However, the high mobility of vehicles and the ad hoc nature of vehicular networks make maintaining stable task offloading performance quite challenging. Especially when the idle computing resources surrounding vehicles are insufficient, there has not been researching to achieve stable offloading performance. Based on this, we consider a proactive strategy that can decrease the events that affect performance stability. First, we propose a mobility prediction model for future network status prediction. Then we design an adaptive task offloading scheme based on proactive adjusting (PATO) to maintain stable task offloading performance. The scheme includes a state processing model and a deep reinforcement learning (DRL)-based task offloading algorithm. Finally, we conduct extensive simulations in various scenarios with insufficient resources to validate the task offloading performance of PATO. Compared with the existing DRL-based algorithm, the simulation results show that PATO can improve the mean offloading utility by 95.4% and the completion ratio by 15.8%.
- Open Access
- 2 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 21699-21711
Multi-Target-Aware Energy Orchestration Modeling for Grid 2.0: A Network Virtualization Approach
Smart Grid– Grid 2.0, is a promising grid scenario due to its low cost, high reliability, and high security. However, the high-intensity complex task interaction and the differentiated Quality of Service (QoS) requirements pose challenges to the energy orchestration of the dynamic and time-varying smart grid. In addition, Network Virtualization (NV) technology is widely utilized to decouple physical network resources. In this work, based on Virtual Network Embedding (VNE), we propose an energy orchestration strategy to better support the operation and maintenance of Grid 2.0. Specifically, a series of constraint indicators for energy orchestration is first formulated. Second, based on the advantages of Deep Reinforcement Learning (DRL), a five-layer architecture energy orchestration model is proposed. It frequently extracts the grid energy matrix as the state input, adopts two-stage actions to reasonably orchestrate the grid energy, and combines the reward mechanism of DRL to effectively interact with the environment. Finally, we conduct experiments in a simulation environment and demonstrate the advantages of the proposed strategy through analysis and comparison.
- 2 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 21699-21711
Open AccessMulti-Target-Aware Energy Orchestration Modeling for Grid 2.0: A Network Virtualization Approach
Smart Grid– Grid 2.0, is a promising grid scenario due to its low cost, high reliability, and high security. However, the high-intensity complex task interaction and the differentiated Quality of Service (QoS) requirements pose challenges to the energy orchestration of the dynamic and time-varying smart grid. In addition, Network Virtualization (NV) technology is widely utilized to decouple physical network resources. In this work, based on Virtual Network Embedding (VNE), we propose an energy orchestration strategy to better support the operation and maintenance of Grid 2.0. Specifically, a series of constraint indicators for energy orchestration is first formulated. Second, based on the advantages of Deep Reinforcement Learning (DRL), a five-layer architecture energy orchestration model is proposed. It frequently extracts the grid energy matrix as the state input, adopts two-stage actions to reasonably orchestrate the grid energy, and combines the reward mechanism of DRL to effectively interact with the environment. Finally, we conduct experiments in a simulation environment and demonstrate the advantages of the proposed strategy through analysis and comparison.