Dehua Zhang
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
Dehua Zhang has a total of 76 co-authors. Here are the top co-authors by number of shared publications:
- Chunbin Qin
Henan University
17 shared publications - Linlin Liang
School of Cyber Engineering, Xidian University, Xian, China
6 shared publications Chunbin Qin
5 shared publications- Jinguang Wang
School of Artificial Intelligence, Henan University, Zhengzhou, China
5 shared publications Mingyu Pang
Henan University
4 shared publicationsSuyang Hou
Henan University
4 shared publications
Latest Publications
- Top Cited Papers
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (4) , 9783-9791
Safe Human-Centric Shared Control for Consumer Service Robots in Smart Home Environments via Observer-Based ADP
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (4) , 9783-9791
Top Cited PapersSafe Human-Centric Shared Control for Consumer Service Robots in Smart Home Environments via Observer-Based ADP
- Open Access
- 15 September 2026
- Published by MDPI AG in Fractal and Fractional
- Vol. 10 (9) , 642
A Review of Image Encryption Driven by Fractional-Order Neural Network Models
Traditional integer-order chaos-based encryption systems often exhibit relatively limited dynamical behavior and may suffer from complexity degradation or short periodic trajectories with finite-precision implementation, thereby weakening their resistance to cryptanalysis. Fractional-Order Neural Networks (FONNs), owing to their inherent memory and hereditary properties as well as their effectively infinite-dimensional dynamics, provide a promising approach to alleviating these limitations. This paper presents a systematic review of Fractional-Order Neural Network (FONN)-based image encryption methods and their recent advances. The novelty of this review lies in its unified classification framework that categorizes FONN-based encryption schemes into three representative architectures: fractional-order Hopfield neural networks, fractional-order memristive neural networks and fractional-order cellular neural networks. It provides a comparison of their dynamical characteristics, encryption mechanisms and security performance. This review provides a coherent theoretical framework and practical reference for further research on FONN-based image encryption.
- 15 September 2026
- Published by MDPI AG in Fractal and Fractional
- Vol. 10 (9) , 642
Open AccessA Review of Image Encryption Driven by Fractional-Order Neural Network Models
Traditional integer-order chaos-based encryption systems often exhibit relatively limited dynamical behavior and may suffer from complexity degradation or short periodic trajectories with finite-precision implementation, thereby weakening their resistance to cryptanalysis. Fractional-Order Neural Networks (FONNs), owing to their inherent memory and hereditary properties as well as their effectively infinite-dimensional dynamics, provide a promising approach to alleviating these limitations. This paper presents a systematic review of Fractional-Order Neural Network (FONN)-based image encryption methods and their recent advances. The novelty of this review lies in its unified classification framework that categorizes FONN-based encryption schemes into three representative architectures: fractional-order Hopfield neural networks, fractional-order memristive neural networks and fractional-order cellular neural networks. It provides a comparison of their dynamical characteristics, encryption mechanisms and security performance. This review provides a coherent theoretical framework and practical reference for further research on FONN-based image encryption.
- Open Access
- 5 August 2026
- Published by MDPI AG in Electronics
- Vol. 15 (15) , 3465
Improved DPT-Hybrid for Monocular Depth Estimation with Geometry-Enhanced Encoding and Structure-Aware Gated Fusion
Monocular depth estimation aims to recover dense 3D scene geometry from a single RGB image and plays an important role in autonomous driving, robotic perception, augmented reality, and 3D reconstruction. Although Transformer-based dense prediction models have achieved strong performance, existing DPT-Hybrid frameworks still suffer from three limitations: insufficient local geometric modeling in shallow stages, inadequate cross-scale fusion for preserving fine structures, and training objectives that only weakly constrain structural consistency. To address these issues, we propose a structure-aware enhanced DPT-Hybrid framework. First, a geometry-enhanced encoder introduces lightweight depth-wise separable convolution branches into shallow Transformer stages to better capture local edge and texture cues while preserving global contextual modeling. Second, a Structure-Aware Cross-Scale Gated Attention Fusion (S-GAF) module is proposed to improve decoder-side feature aggregation by jointly modeling channel-wise and spatial importance with an auxiliary RGB-gradient input. Third, joint structure–geometric consistency loss combines scale-invariant logarithmic loss, gradient consistency loss, and edge-focused loss to improve pixel-level accuracy, geometric plausibility, and boundary sharpness. Experiments on NYUv2 and KITTI demonstrate that the proposed method achieves lower single-run error metrics than the controlled DPT-Hybrid baseline under the evaluated settings. On NYUv2, our method achieves an absolute relative error (AbsRel) of 0.099 and an RMSE of 0.334. On KITTI, it achieves an AbsRel of 0.058 and an RMSE of 2.455. The proposed method introduces only modest additional complexity while producing more accurate and structurally sharper depth predictions.
- 5 August 2026
- Published by MDPI AG in Electronics
- Vol. 15 (15) , 3465
Open AccessImproved DPT-Hybrid for Monocular Depth Estimation with Geometry-Enhanced Encoding and Structure-Aware Gated Fusion
Monocular depth estimation aims to recover dense 3D scene geometry from a single RGB image and plays an important role in autonomous driving, robotic perception, augmented reality, and 3D reconstruction. Although Transformer-based dense prediction models have achieved strong performance, existing DPT-Hybrid frameworks still suffer from three limitations: insufficient local geometric modeling in shallow stages, inadequate cross-scale fusion for preserving fine structures, and training objectives that only weakly constrain structural consistency. To address these issues, we propose a structure-aware enhanced DPT-Hybrid framework. First, a geometry-enhanced encoder introduces lightweight depth-wise separable convolution branches into shallow Transformer stages to better capture local edge and texture cues while preserving global contextual modeling. Second, a Structure-Aware Cross-Scale Gated Attention Fusion (S-GAF) module is proposed to improve decoder-side feature aggregation by jointly modeling channel-wise and spatial importance with an auxiliary RGB-gradient input. Third, joint structure–geometric consistency loss combines scale-invariant logarithmic loss, gradient consistency loss, and edge-focused loss to improve pixel-level accuracy, geometric plausibility, and boundary sharpness. Experiments on NYUv2 and KITTI demonstrate that the proposed method achieves lower single-run error metrics than the controlled DPT-Hybrid baseline under the evaluated settings. On NYUv2, our method achieves an absolute relative error (AbsRel) of 0.099 and an RMSE of 0.334. On KITTI, it achieves an AbsRel of 0.058 and an RMSE of 2.455. The proposed method introduces only modest additional complexity while producing more accurate and structurally sharper depth predictions.
- Top Cited Papers
- 30 July 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 333
Scale-spectrum-state collaborative fusion network for real-time extreme small object detection in intelligent sports analytics
- 30 July 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 333
Top Cited PapersScale-spectrum-state collaborative fusion network for real-time extreme small object detection in intelligent sports analytics
- Top Cited Papers
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (13) , 28694-28706
Sec-GNN-Driven Joint Multidimensional Anti-Eavesdropping Optimization for Secure UAV-Satellite Communications
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (13) , 28694-28706
Top Cited PapersSec-GNN-Driven Joint Multidimensional Anti-Eavesdropping Optimization for Secure UAV-Satellite Communications
- Top Cited Papers
- 7 March 2026
- Published by Elsevier in ISA Transactions
- Vol. 172, 21-32
Dynamic event-triggered approximate optimal consensus control for unknown nonlinear multi-agent systems via adaptive dynamic programming
- 7 March 2026
- Published by Elsevier in ISA Transactions
- Vol. 172, 21-32
Top Cited PapersDynamic event-triggered approximate optimal consensus control for unknown nonlinear multi-agent systems via adaptive dynamic programming
- Top Cited Papers
- 19 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 56 (6) , 3733-3744
Dynamic Event-Triggered Control for Human–Machine Cooperative Systems Based on Dynamic Authority Allocation
This article addresses the challenging problem of constrained optimal control for human–machine systems subject to external disturbances and the bounded rationality of the human operator. To this end, a novel game-theoretic framework is proposed. Unlike monolithic game formulations, the framework uniquely disaggregates the control problem by transforming it into a multifaceted game via logarithmic barrier functions (BFs): it models human–machine cooperation as a positive-sum game oriented toward shared objectives, and disturbance rejection as a zero-sum game tailored for robustness enhancement. To capture the nonideal human decision-making, we integrate the level- $k$ reasoning framework to model the operator’s bounded cognitive dynamics. The corresponding coupled Hamilton–Jacobi–Isaacs (HJI) equations for this human–machine game are derived, and critically, a rigorous proof of global asymptotic stability (GAS) for the transformed system is provided, establishing a solid theoretical foundation. For online implementation without requiring prior knowledge of the system dynamics, we develop a resource-efficient learning architecture based on the adaptive dynamic programming (ADP) and a novel dynamic event-triggered mechanism (DETM). A key feature of this architecture is a fuzzy logic-based module for dynamic authority allocation, which adaptively adjusts control sharing in real time. Rigorous analysis demonstrates that all signals in the closed-loop system are uniformly ultimately bounded and that Zeno behavior is precluded. Simulation results are presented to validate the effectiveness and superiority of the proposed control strategy.
- 19 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 56 (6) , 3733-3744
Top Cited PapersDynamic Event-Triggered Control for Human–Machine Cooperative Systems Based on Dynamic Authority Allocation
This article addresses the challenging problem of constrained optimal control for human–machine systems subject to external disturbances and the bounded rationality of the human operator. To this end, a novel game-theoretic framework is proposed. Unlike monolithic game formulations, the framework uniquely disaggregates the control problem by transforming it into a multifaceted game via logarithmic barrier functions (BFs): it models human–machine cooperation as a positive-sum game oriented toward shared objectives, and disturbance rejection as a zero-sum game tailored for robustness enhancement. To capture the nonideal human decision-making, we integrate the level- $k$ reasoning framework to model the operator’s bounded cognitive dynamics. The corresponding coupled Hamilton–Jacobi–Isaacs (HJI) equations for this human–machine game are derived, and critically, a rigorous proof of global asymptotic stability (GAS) for the transformed system is provided, establishing a solid theoretical foundation. For online implementation without requiring prior knowledge of the system dynamics, we develop a resource-efficient learning architecture based on the adaptive dynamic programming (ADP) and a novel dynamic event-triggered mechanism (DETM). A key feature of this architecture is a fuzzy logic-based module for dynamic authority allocation, which adaptively adjusts control sharing in real time. Rigorous analysis demonstrates that all signals in the closed-loop system are uniformly ultimately bounded and that Zeno behavior is precluded. Simulation results are presented to validate the effectiveness and superiority of the proposed control strategy.
- Top Cited Papers
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (1) , 455-465
Event-Triggered Zero-Sum Game for Input Saturated Nonlinear Systems With State Constraints
In safety-critical applications, saturated nonlinear systems often face significant control challenges due to the simultaneous presence of strict state constraints, input saturation, and uncertain external disturbances. To address these challenges, this paper proposes a novel event-triggered optimal control (ETOC) method. The core of our approach lies in a unified framework that handles all three issues simultaneously. We first introduce a smooth mapping function to transform the original system with state constraints into an equivalent unconstrained system. Subsequently, the control problem is formulated as a two-player zero-sum game (ZSG), where one player is the controller designed to handle input saturation and the other is the worst-case external disturbance. This unified formulation ensures that the system strictly adheres to the state constraints. Unlike traditional methods that typically require formulating and solving complex modified event-triggered Hamilton-Jacobi-Bellman (HJB) equations incorporating triggering errors, our approach approximates the optimal policy for the transformed system and implements it via an event-triggered mechanism. By utilizing a single-critic neural network (NN) and incorporating an experience replay mechanism, the algorithm’s learning stability is enhanced, and the restrictive persistence of excitation (PE) condition is relaxed. Theoretical analysis and simulations confirm that our ETOC strategy guarantees optimal performance, strict adherence to state constraints, and a significant reduction in communication burden, showcasing its efficiency and practicality.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (1) , 455-465
Top Cited PapersEvent-Triggered Zero-Sum Game for Input Saturated Nonlinear Systems With State Constraints
In safety-critical applications, saturated nonlinear systems often face significant control challenges due to the simultaneous presence of strict state constraints, input saturation, and uncertain external disturbances. To address these challenges, this paper proposes a novel event-triggered optimal control (ETOC) method. The core of our approach lies in a unified framework that handles all three issues simultaneously. We first introduce a smooth mapping function to transform the original system with state constraints into an equivalent unconstrained system. Subsequently, the control problem is formulated as a two-player zero-sum game (ZSG), where one player is the controller designed to handle input saturation and the other is the worst-case external disturbance. This unified formulation ensures that the system strictly adheres to the state constraints. Unlike traditional methods that typically require formulating and solving complex modified event-triggered Hamilton-Jacobi-Bellman (HJB) equations incorporating triggering errors, our approach approximates the optimal policy for the transformed system and implements it via an event-triggered mechanism. By utilizing a single-critic neural network (NN) and incorporating an experience replay mechanism, the algorithm’s learning stability is enhanced, and the restrictive persistence of excitation (PE) condition is relaxed. Theoretical analysis and simulations confirm that our ETOC strategy guarantees optimal performance, strict adherence to state constraints, and a significant reduction in communication burden, showcasing its efficiency and practicality.