MSMing Sun

Ming Sun

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Ming Sun has a total of 24 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Top Cited Papers
    Open Access

    State-Aware Resource Allocation for V2X Communications

    by Ming Sun, Jinqing Xu, Jiaying Wang

    Vehicle-to-Everything (V2X) has become a key technology for addressing intelligent transportation challenges. Improving spectrum utilization and mitigating multi-user interference among V2X links are currently the primary focuses of research efforts. However, the time-varying nature of channel resources and the dynamic vehicular environment pose significant challenges to achieving high spectral efficiency and low interference. Numerous studies have demonstrated the effectiveness of deep reinforcement learning (DRL) in distributed resource allocation for vehicular networks. Nevertheless, in conventional distributed DRL frameworks, the independence of agent decisions often weakens cooperation among agents, thereby limiting the overall performance potential of the algorithms. To address this limitation, this paper proposes a state-aware communication resource allocation algorithm for vehicular networks. The proposed approach enhances the representation capability of observable data by expanding the state space, thus improving the utilization of available observations. Additionally, a conditional attention mechanism is introduced to strengthen the model’s perception of environmental dynamics. These innovative improvements significantly enhance each agent’s awareness of the environment and promote effective collaboration among agents. Simulation results verify that the proposed algorithm effectively improves agents’ environmental perception and inter-agent cooperation, leading to superior performance in complex and dynamic V2X scenarios.


  • Open Access

    Spatial and Spectral Structure-Aware Mamba Network for Hyperspectral Image Classification

    by Jie Zhang, Ming Sun, Sheng Chang

    Recently, a network based on selective state space models (SSMs), Mamba, has emerged as a research focus in hyperspectral image (HSI) classification due to its linear computational complexity and strong long-range dependency modeling capability. Originally designed for 1D causal sequence modeling, Mamba is challenging for HSI tasks that require simultaneous awareness of spatial and spectral structures. Current Mamba-based HSI classification methods typically convert spatial structures into 1D sequences and employ various scanning patterns to capture spatial dependencies. However, these approaches inevitably disrupt spatial structures, leading to ineffective modeling of complex spatial relationships and increased computational costs due to elongated scanning paths. Moreover, the lack of neighborhood spectral information utilization fails to mitigate the impact of spatial variability on classification performance. To address these limitations, we propose a novel model, Dual-Aware Discriminative Fusion Mamba (DADFMamba), which is simultaneously aware of spatial-spectral structures and adaptively integrates discriminative features. Specifically, we design a Spatial-Structure-Aware Fusion Module (SSAFM) to directly establish spatial neighborhood connectivity in the state space, preserving structural integrity. Then, we introduce a Spectral-Neighbor-Group Fusion Module (SNGFM). It enhances target spectral features by leveraging neighborhood spectral information before partitioning them into multiple spectral groups to explore relations across these groups. Finally, we introduce a Feature Fusion Discriminator (FFD) to discriminate the importance of spatial and spectral features, enabling adaptive feature fusion. Extensive experiments on four benchmark HSI datasets demonstrate that DADFMamba outperforms state-of-the-art deep learning models in classification accuracy while maintaining low computational costs and parameter efficiency. Notably, it achieves superior performance with only 30 training samples per class, highlighting its data efficiency. Our study reveals the great potential of Mamba in HSI classification and provides valuable insights for future research.


  • Open Access

    Channel and Power Allocation for Multi-Cell NOMA Using Multi-Agent Deep Reinforcement Learning and Unsupervised Learning

    by Ming Sun, Yihe Zhong, Xiaoou He, Jie Zhang

    Among the 5G and anticipated 6G technologies, non-orthogonal multiple access (NOMA) has attracted considerable attention due to its notable advantages in data throughput. Nevertheless, it is challenging to find the near-optimal allocation of the channel and power resources to maximize the performance of the multi-cell NOMA system. In addition, due to the complex and dynamically changing wireless communication environment and the lack of the near-optimal labels, conventional supervised learning methods cannot be directly applied. To address these challenges, this paper proposes a framework of MDRL-UL that integrates the multi-agent deep reinforcement learning with the unsupervised learning to allocate the channel and power resources in a near-optimal manner. In the framework, a multi-agent deep reinforcement learning neural network (MDRLNN) is proposed for channel allocation, while an attention-based unsupervised learning neural network (ULNN) is proposed for power allocation. Furthermore, the joint action (JA) derived from the MDRLNN for channel allocation is used as a representation to be fed into the ULNN for power allocation. In order to maximize the energy efficiency of the multi-cell NOMA system, the expectation of the energy efficiency is used to train both the MDRLNN and the ULNN. Simulation results indicate that the proposed MDRL-UL can achieve higher energy efficiency and transmission rates than other algorithms.


  • Open Access

    Bidirectional Mamba with Dual-Branch Feature Extraction for Hyperspectral Image Classification

    by Ming Sun, Jie Zhang, Xiaoou He, Yihe Zhong

    The hyperspectral image (HSI) classification task is widely used in remote sensing image analysis. The HSI classification methods based on convolutional neural networks (CNNs) have greatly improved the classification performance. However, they cannot well utilize the sequential properties of spectral features and face the challenge of increasing computational cost with the increase in network depth. To address these shortcomings, this paper proposes a novel network with a CNN-Mamba architecture, called DBMamba, which uses a bidirectional Mamba to process spectral feature sequences at a linear computational cost. In the DBMamba, principal component analysis (PCA) is first used to extract the main features of the data. Then, a dual-branch CNN structure, with the fused features from spectral–spatial features by 3D-CNN and spatial features by 2D-CNN, is used to extract shallow spectral–spatial features. Finally, a bidirectional Mamba is used to effectively capture global contextual information in features and significantly enhance the extraction of spectral features. Experimental results on the Indian Pines, Salinas, and Pavia University datasets demonstrate that the classification performance surpasses that of many cutting-edge methods, improving by 1.04%, 0.15%, and 0.09%, respectively, over the competing SSFTT method. The research in this paper enhances the existing knowledge on HSI classification and provides valuable insights for future research in this field.


  • Open Access

    Knowledge Transfer within Enterprises from the Perspective of Innovation Quality Management: A Decision Analysis Based on the Stackelberg Game

    by Shumei Wang, Ming Sun, Yaoqun Xu

    It is of great significance to study the effectiveness of knowledge transfer between the new and the veteran employees within enterprises for promoting sustainable innovation from the perspective of innovation quality management. However, few studies have examined the impact of innovation quality on the effectiveness of knowledge transfer between the new and veteran employees. In addition, knowledge of how reward and punishment incentives affect the effectiveness of knowledge transfer in innovation quality management remains insufficient. Since the amount of knowledge transfer is an important aspect of the effectiveness of knowledge transfer, this paper constructs a Stackelberg game model with an innovation-quality-oriented threshold of the knowledge transfer amount and investigates the amount of knowledge transfer between new and veteran employees in the collaborative innovation of research and development (R&D) projects within enterprises. A case study was used to reveal that the innovation-quality-oriented threshold for the knowledge transfer amount can effectively promote the amount of knowledge transfer between the new and the veteran employees in collaborative innovation. Moreover, reward is more effective than punishment in promoting the amount of knowledge transfer to exceed the innovation-quality-oriented threshold. This study enriches the theories of knowledge transfer games under quality management. By virtue of end-to-end project management strategies, modern multimedia technologies, and reward incentives this study can be used to conduct quality control during project execution, to promote knowledge retention in R&D projects, the innovation quality of projects, and the achievement of the Sustainable Development Goals (SDGs). The research methodology employed in this paper was limited to a case study, and the data utilized are not empirical data.


  • Open Access

    Joint DDPG and Unsupervised Learning for Channel Allocation and Power Control in Centralized Wireless Cellular Networks

    by Ming Sun, Erzhuang Mei, Shumei Wang, Yanhui Jin

    In order to solve the resource allocation problem in scenarios of centralized wireless cellular communication with multiple cells, users and channels, a novel resource allocation algorithm based on joint Deep Deterministic Policy Gradient (DDPG) reinforcement learning and unsupervised learning is proposed. Firstly, the proposed algorithm builds a channel allocation deep neural network based on DDPG to provide an optimized channel allocation scheme. Secondly, the proposed algorithm constructs a power control deep neural network based on unsupervised learning to provide an optimized power control scheme. In order to make the unsupervised learning have perceptions on dynamic wireless environments, the experience replay is executed twice to train the channel allocation deep neural network with the DDPG reinforcement learning and the power control deep neural network with the unsupervised learning, respectively. Because the proposed joint algorithm combines the dynamic perception ability of the DDPG reinforcement learning and the continuous optimization ability of unsupervised learning, the energy efficiency can be maximized effectively. Simulation results show that the proposed algorithm outperforms other algorithms in terms of energy efficiency and transmit rate in time-varying dynamic environments.


  • Open Access

    Joint Deep Reinforcement Learning and Unsupervised Learning for Channel Selection and Power Control in D2D Networks

    by Ming Sun, Yanhui Jin, Shumei Wang, Erzhuang Mei

    Device-to-device (D2D) technology enables direct communication between devices, which can effectively solve the problem of insufficient spectrum resources in 5G communication technology. Since the channels are shared among multiple D2D user pairs, it may lead to serious interference between D2D user pairs. In order to reduce interference, effectively increase network capacity, and improve wireless spectrum utilization, this paper proposed a distributed resource allocation algorithm with the joint of a deep Q network (DQN) and an unsupervised learning network. Firstly, a DQN algorithm was constructed to solve the channel allocation in the dynamic and unknown environment in a distributed manner. Then, a deep power control neural network with the unsupervised learning strategy was constructed to output an optimized channel power control scheme to maximize the spectrum transmit sum-rate through the corresponding constraint processing. As opposed to traditional centralized approaches that require the collection of instantaneous global network information, the algorithm proposed in this paper used each transmitter as a learning agent to make channel selection and power control through a small amount of state information collected locally. The simulation results showed that the proposed algorithm was more effective in increasing the convergence speed and maximizing the transmit sum-rate than other traditional centralized and distributed algorithms.


  • Open Access

    A New Bearing Fault Diagnosis Method Based on Capsule Network and Markov Transition Field/Gramian Angular Field

    by Bin Han, Hui Zhang, Ming Sun, Fengtong Wu

    Compared to time-consuming and unreliable manual analysis, intelligent fault diagnosis techniques using deep learning models can improve the accuracy of intelligent fault diagnosis with their multi-layer nonlinear mapping capabilities. This paper proposes a model to perform fault diagnosis and classification by using a time series of vibration sensor data as the input. The model encodes the raw vibration signal into a two-dimensional image and performs feature extraction and classification by a deep convolutional neural network or improved capsule network. A fault diagnosis technique based on the Gramian Angular Field (GAF), the Markov Transition Field (MTF), and the Capsule Network is proposed. Experiments conducted on a bearing failure dataset from Case Western Reserve University investigated the impact of two coding methods and different network structures on the diagnosis accuracy. The results show that the GAF technique retains more complete fault characteristics, while the MTF technique contains a small number of fault characteristics but more dynamic characteristics. Therefore, the proposed method incorporates GAF images and MTF images as a dual-channel image input to the capsule network, enabling the network to obtain a more complete fault signature. Multiple sets of experiments were conducted on the bearing fault dataset at Case Western Reserve University, and the Capsule Network in the proposed model has an advantage over other convolutional neural networks and performs well in the comparison of fault diagnosis methods proposed by other researchers.


  • Open Access

    Novel Bee Colony Optimization with Update Quantities for OFDMA Resource Allocation

    by Ming Sun, Yujing Huang, Shumei Wang, Yaoqun Xu

    In recent years, large usage of wireless networks puts forward challenge to the utilization of spectrum resources, and it is significant to improve the spectrum utilization and the system sum data rates in the premise of fairness. However, the existing algorithms have drawbacks in efficiency to maximize the sum data rates of orthogonal frequency division multiple access (OFDMA) systems in the premise of fairness threshold. To address the issue, a novel artificial bee colony algorithm with update quantities of nectar sources is proposed for OFDMA resource allocation in this paper. Firstly, the population of nectar sources is divided into several groups, and a different update quantity of nectar sources is set for each group. Secondly, based on the update quantities of nectar sources set for these groups, nectar sources are initialized by a greedy subcarrier allocation method. Thirdly, neighborhood searches and updates are performed on dimensions of nectar sources corresponding to the preset update quantities. The proposed algorithm can not only make the initialized nectar sources maintain high levels of fairness through the greedy subcarrier allocation but also use the preset update quantities to reduce dimensions of the nectar sources to be optimized by the artificial bee colony algorithm, thereby making full use of both the local optimization of the greedy method and the global optimization of the artificial bee colony algorithm. The simulation results show that, just in the equal‐power subcarrier allocation stage, the proposed algorithm can achieve the required fairness threshold and effectively improve the system sum data rates.


  • Open Access

    Hybrid OFDMA Resource Allocation Scheme for Ensuring Required Level of Proportional Fairness

    by Ming Sun, Kangle Zhai, Wei Cao, Ying Wang, Yaoqun Xu

    The new-generation wireless communication networks are envisioned to offer higher sum data rates along with the required level of fairness. Previous works tend to suffer from a decayed performance as subcarriers become relatively insufficient in allocation to users. To maximize the sum data rates and ensure the required level of proportional fairness, this paper presents a hybrid OFDMA resource allocation scheme which uses Hungarian algorithm combined with a greedy method for subcarrier allocation and uses bee colony optimization for power allocation. The proposed subcarrier allocation scheme can make full use of advantages of both globally optimal Hungarian algorithm in enhancing sum data rates and locally optimal greedy method in maintaining a reasonable fairness level and can make Hungarian algorithm work in a searching mode for further improvement of sum data rates and fairness. The proposed power allocation scheme can converge to the required level of proportional fairness but with higher sum data rates if the subcarrier allocation does not achieve the required fairness. Simulation results show that the proposed scheme can obtain the required level of proportional fairness but with higher sum data rates even if subcarriers are relatively insufficient in allocation to users. Complexity analysis shows the proposed method has moderate complexity.