FZFengjun Zhou

Fengjun Zhou

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Fengjun Zhou has a total of 28 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Multi-HypPre: a multi-source data fusion framework for drug-induced hypertension prediction

    by Xueqiang Gao, Chen Ge, Kui Lu, Guodong Wang, Fengjun Zhou, Yilei Wang, Huiyu Zhou, Xiaoying Wang, Wei Dai

  • Reinforcement Learning-Based Adaptive Control Optimization for Nonlinear Systems

    by Xueqiang Gao, Guodong Wang, Lei Chen, Xiu Huang, Fengjun Zhou

    This paper introduces a novel approach to adaptive control optimization for nonlinear systems using reinforcement learning (RL). The proposed algorithm, RL-ACOS (Reinforcement Learning-based Adaptive Control Optimization System), employs RL to adjust control strategies dynamically, responding to real-time environmental feedback. By merging RL with adaptive control techniques, this algorithm optimizes system performance, enhances stability, and bolsters resilience to nonlinear disturbances and uncertainties. Through extensive simulation experiments, we demonstrate that RL-ACOS outperforms conventional control methods in terms of both convergence speed and overall system efficiency, offering a promising solution for complex nonlinear systems in real-world applications.

  • Open Access

    QL-HIT2F: A Q-Learning-Aided Adaptive Fuzzy Path Planning Algorithm with Enhanced Obstacle Avoidance

    by Nana Zhou, Fengjun Zhou, Changming Li, Jianning Zhong

    There has been significant interest in solving robot path planning problems using fuzzy logic-based methods. Recently, the Genetic Algorithm-based Hierarchical Interval Type-2 Fuzzy (GA-HIT2F) system has been introduced as a novel planner in this domain. However, this method lacks adaptability to changes in target points, and insufficient flexibility can lead to planning failures in local minimum traps, making it difficult to apply to complex scenarios. In this paper, we identify the limitations of the original GA-HIT2F approach and propose an enhanced Q-Learning-aided Adaptive Hierarchical Interval Type-2 Fuzzy (QL-HIT2F) algorithm for path planning. The proposed planner incorporates reinforcement learning to improve a robot’s capability to avoid collisions with special obstacles. Additionally, the average obstacle orientation (AOO) is introduced to optimize the robot’s angular adjustments. Two supplementary robot parameters are integrated into the reinforcement learning action space, along with fuzzy membership parameters. The training process also introduces the concepts of meta-map and sub-training. Simulation results from a series of path planning experiments validate the feasibility and effectiveness of the proposed QL-HIT2F approach.

  • Open Access

    Security Design for Data Distribution Service Based on Secure Onboard Communication

    by Fengjun Zhou, Gang Wang, Qun Wang, Yuxing Wang, Jun Wang

    Data Distribution Service (DDS) middleware plays a critical role in intelligent connected vehicle (ICV) communication; however, its existing security mechanisms often compromise real-time performance. This study proposes a hierarchical dynamic security protection architecture that achieves deep integration of Secure Onboard Communication(SecOC) and DDS security through protocol extension, policy mapping, and resource degradation. The core innovations are as follows: 1) Dynamic Security Policy Adaptation across Protocol Layers: By extending the Real-Time Publish-Subscribe protocol of DDS, SecOC security fields (Freshness Value + Message Authentication Code) are embedded into the message structure, and a dynamic mapping mechanism is established between the topic-level Quality of Service policies and the Protocol Data Unit-level security parameters is established. Security levels are defined based on Extensible Markup Language configuration files to achieve automated policy loading from application layer encryption to transport layer authentication, thereby solving the problem of consistency in security policies between heterogeneous protocol stacks. 2) Lightweight Security Grading Method: A security grading method: A security grading framework was constructed based on the sensitivity of DDS topics. This method balances security strength and computational overhead through differentiated security processing while considering the ISO 26262 functional safety standard. Compared with the original security plugin, experimental verification in typical communication scenarios demonstrates its superiority in jitter reduction by 35%, latency reduction by 50%, and throughput improvement of approximately 12%, thereby effectively fulfilling the dual requirements of ICV for robust security and high real-time performance.


  • A Transformer-Based Method for Few-Shot Image Recognition in Federated Learning Environments

    by Guodong Wang, Qianqian Li, Xinyi Lin, Zixuan Wang, Fengjun Zhou, Qun Wang, Weiguo Lai, Xueqiang Gao

    Few-shot image recognition in federated learning (FL) environments presents unique challenges due to data heterogeneity, limited local samples, and strict privacy constraints. To address these issues, we propose a novel Transformer-based approach that enables effective visual recognition with minimal labeled data on distributed clients. The proposed model introduces an adaptive multi-scale window attention mechanism to enhance feature extraction under sparse data conditions. Furthermore, a collaborative attention strategy is developed to enable privacy-preserving semantic knowledge exchange across clients. To alleviate the class imbalance and overfitting problems common in few-shot settings, we design an improved Focal Loss function guided by class-wise entropy. Additionally, a progressive knowledge distillation framework is incorporated to maintain model consistency across training rounds. Extensive experiments conducted on multiple few-shot image benchmarks under federated settings demonstrate that our method significantly outperforms existing baselines in terms of accuracy, communication efficiency, and generalization.