Zijian Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zijian Wang has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Lei Chen
School of Computer Science and Information Engineering and the Key Laboratory of Knowledge Engineering with Big Data, Ministry of Education, Hefei University of Technology, Hefei, China
2 shared publications - Na Xia
Hefei University of Technology
2 shared publications - Sizhou Wei
School of Computer Science and Information Engineering, Hefei University of Technology, Hefei, China
2 shared publications - Yin Wang
School of Computer Science and Information Engineering, Hefei University of Technology, Hefei, China
2 shared publications - Yutao Yin
School of Computer and Information Engineering, Hefei University of Technology, Hefei, China
2 shared publications - Zongyun Gu
School of Medical Information Engineering, Anhui University of Chinese Medicine, Hefei 230012, P. R. China
2 shared publications
Latest Publications
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Green Communications and Networking
- Vol. 10, 218-235
CoMARS: Coordinated Multi-Agent Rollout Strategy for Distributed Channel Access in Clustered IoT Networks
The proliferation of Machine-to-Machine (M2M) communications in Internet of Things (IoT) networks intensifies interference management and resource allocation challenges for clustered devices. Existing solutions face limitations in scalability and fairness, particularly under dynamic conditions. We propose CoMARS, a bio-inspired framework integrating artificial immune systems with multi-agent reinforcement learning, designed to address these challenges through coordinated two-tier optimization. The architecture employs immune-inspired mechanisms at the device layer for autonomous cluster maintenance and interference-aware power control. At the network layer, a predictive rollout algorithm enables cluster heads to develop coordinated transmission strategies through model-based reinforcement learning. OMNeT++ simulations demonstrate CoMARS’s superior performance, achieving 5.71 Mbps/J energy efficiency with significantly improved throughput and lower latency compared to benchmark protocols. The framework maintains stable performance during topology changes and exhibits accelerated convergence under saturated traffic, proving effective for dynamic IoT environments.
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Green Communications and Networking
- Vol. 10, 218-235
CoMARS: Coordinated Multi-Agent Rollout Strategy for Distributed Channel Access in Clustered IoT Networks
The proliferation of Machine-to-Machine (M2M) communications in Internet of Things (IoT) networks intensifies interference management and resource allocation challenges for clustered devices. Existing solutions face limitations in scalability and fairness, particularly under dynamic conditions. We propose CoMARS, a bio-inspired framework integrating artificial immune systems with multi-agent reinforcement learning, designed to address these challenges through coordinated two-tier optimization. The architecture employs immune-inspired mechanisms at the device layer for autonomous cluster maintenance and interference-aware power control. At the network layer, a predictive rollout algorithm enables cluster heads to develop coordinated transmission strategies through model-based reinforcement learning. OMNeT++ simulations demonstrate CoMARS’s superior performance, achieving 5.71 Mbps/J energy efficiency with significantly improved throughput and lower latency compared to benchmark protocols. The framework maintains stable performance during topology changes and exhibits accelerated convergence under saturated traffic, proving effective for dynamic IoT environments.
- 24 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 22 (1) , 846-859
Flow Adjustment and Scale-Free Reliable Topology Control for Underwater Acoustic Sensor Networks
To improve the reliability and efficiency of underwater acoustic sensor networks (UASNs) under limited node energy and network attacks, a joint scheme of enhancing robustness and extending network lifetime is proposed in this paper. Scale-free networks exhibit strong robustness against random failures, and because of the lower number of link connections, nodes consume energy more slowly. We first introduce a new scale-free topology evolution model that adjusts the initial number of nodes according to the characteristics of UASNs. This model incorporates factors such as flow load, energy consumption, and distance into the preferential attachment mechanism, balancing the network load and enhancing its resistance to attacks. Further, based on this topology, we propose a network flow adjustment algorithm that features the joint selection of paths and corresponding power levels. Energy consumption is balanced among nodes in proportion to their residual energy, rather than by minimizing the absolute consumed power. Numerical simulations show that different topologies significantly affect network lifetime, which is the longest when the number of links added is 2. The network lifetime of the proposed scheme surpasses the state-of-the-art schemes, such as ETFLA, Initial-BA, and VODA, by up to 37.18%, 46.23%, and 128.80%, respectively.
- 24 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 22 (1) , 846-859
Flow Adjustment and Scale-Free Reliable Topology Control for Underwater Acoustic Sensor Networks
To improve the reliability and efficiency of underwater acoustic sensor networks (UASNs) under limited node energy and network attacks, a joint scheme of enhancing robustness and extending network lifetime is proposed in this paper. Scale-free networks exhibit strong robustness against random failures, and because of the lower number of link connections, nodes consume energy more slowly. We first introduce a new scale-free topology evolution model that adjusts the initial number of nodes according to the characteristics of UASNs. This model incorporates factors such as flow load, energy consumption, and distance into the preferential attachment mechanism, balancing the network load and enhancing its resistance to attacks. Further, based on this topology, we propose a network flow adjustment algorithm that features the joint selection of paths and corresponding power levels. Energy consumption is balanced among nodes in proportion to their residual energy, rather than by minimizing the absolute consumed power. Numerical simulations show that different topologies significantly affect network lifetime, which is the longest when the number of links added is 2. The network lifetime of the proposed scheme surpasses the state-of-the-art schemes, such as ETFLA, Initial-BA, and VODA, by up to 37.18%, 46.23%, and 128.80%, respectively.
- Top Cited PapersOpen Access
- 3 January 2023
- Published by Wiley in Computational Intelligence and Neuroscience
- Vol. 2023 (1) , 1305583
Classification of Diabetic Retinopathy Severity in Fundus Images Using the Vision Transformer and Residual Attention
Diabetic retinopathy (DR) is a common retinal vascular disease, which can cause severe visual impairment. It is of great clinical significance to use fundus images for intelligent diagnosis of DR. In this paper, an intelligent DR classification model of fundus images is proposed. This method can detect all the five stages of DR, including of no DR, mild, moderate, severe, and proliferative. This model is composed of two key modules. FEB, feature extraction block, is mainly used for feature extraction of fundus images, and GPB, grading prediction block, is used to classify the five stages of DR. The transformer in the FEB has more fine‐grained attention that can pay more attention to retinal hemorrhage and exudate areas. The residual attention in the GPB can effectively capture different spatial regions occupied by different classes of objects. Comprehensive experiments on DDR datasets well demonstrate the superiority of our method, and compared with the benchmark method, our method has achieved competitive performance.
- 3 January 2023
- Published by Wiley in Computational Intelligence and Neuroscience
- Vol. 2023 (1) , 1305583
Top Cited PapersOpen AccessClassification of Diabetic Retinopathy Severity in Fundus Images Using the Vision Transformer and Residual Attention
Diabetic retinopathy (DR) is a common retinal vascular disease, which can cause severe visual impairment. It is of great clinical significance to use fundus images for intelligent diagnosis of DR. In this paper, an intelligent DR classification model of fundus images is proposed. This method can detect all the five stages of DR, including of no DR, mild, moderate, severe, and proliferative. This model is composed of two key modules. FEB, feature extraction block, is mainly used for feature extraction of fundus images, and GPB, grading prediction block, is used to classify the five stages of DR. The transformer in the FEB has more fine‐grained attention that can pay more attention to retinal hemorrhage and exudate areas. The residual attention in the GPB can effectively capture different spatial regions occupied by different classes of objects. Comprehensive experiments on DDR datasets well demonstrate the superiority of our method, and compared with the benchmark method, our method has achieved competitive performance.
- Open Access
- 1 January 2022
- Published by Wiley in Complexity
- Vol. 2022 (1)
An Efficient CNN for Hand X‐Ray Overall Scoring of Rheumatoid Arthritis
Rheumatoid arthritis (RA) is a progressive systemic autoimmune disease characterized by inflammation of the joints and surrounding tissues, which seriously affects the life of patients. The Sharp/van der Heijde method has been widely used in clinical evaluation for the RA disease. However, this manual method is time‐consuming and laborious. Even if two radiologists evaluate a specific location, their subjective evaluation may lead to low inter‐rater reliability. Here, we developed an efficient model powered by deep convolutional neural networks to solve these problems and automated the overall scoring on hand X‐rays. The depthwise separable (Dwise) convolution technique is used based on ResNet‐50 due to the high resolution of hand X‐rays. An inverted residual block is introduced to devise a ResNet‐Dwise50 model to enhance the efficiency of the model. The model was trained and tested using bilateral posteroanterior (two‐handed, side by side) images of 3818 patients. The experiment results show the ResNet‐Dwise50 model achieved an MAE of 14.90 and RMSE of 22.01 while ensuring high efficiency. There was no statistical difference between the average scores given by two experienced radiologists and predicted scores from our model.
- 1 January 2022
- Published by Wiley in Complexity
- Vol. 2022 (1)
Open AccessAn Efficient CNN for Hand X‐Ray Overall Scoring of Rheumatoid Arthritis
Rheumatoid arthritis (RA) is a progressive systemic autoimmune disease characterized by inflammation of the joints and surrounding tissues, which seriously affects the life of patients. The Sharp/van der Heijde method has been widely used in clinical evaluation for the RA disease. However, this manual method is time‐consuming and laborious. Even if two radiologists evaluate a specific location, their subjective evaluation may lead to low inter‐rater reliability. Here, we developed an efficient model powered by deep convolutional neural networks to solve these problems and automated the overall scoring on hand X‐rays. The depthwise separable (Dwise) convolution technique is used based on ResNet‐50 due to the high resolution of hand X‐rays. An inverted residual block is introduced to devise a ResNet‐Dwise50 model to enhance the efficiency of the model. The model was trained and tested using bilateral posteroanterior (two‐handed, side by side) images of 3818 patients. The experiment results show the ResNet‐Dwise50 model achieved an MAE of 14.90 and RMSE of 22.01 while ensuring high efficiency. There was no statistical difference between the average scores given by two experienced radiologists and predicted scores from our model.