Zhen He
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zhen He has a total of 32 co-authors. Here are the top co-authors by number of shared publications:
- Di Zhou
School of Astronautics, Harbin Institute of Technology, Harbin, China
5 shared publications - Qiang Wang
Department of Control Science and Engineering, Harbin Institute of Technology
3 shared publications - Yu Bai
School of Astronautics, Harbin Institute of Technology, Harbin 150001, China
3 shared publications - Chao Xu
Research Institute of Intelligent Control and Systems, Harbin Institute of Technology, Harbin, China
2 shared publications - Huijun Gao
Research Institute of Intelligent Control and Systems, Harbin Institute of Technology
2 shared publications - Leiyang Xu
Department of Control Science and Engineering, Harbin Institute of Technology, Harbin 150001, China
2 shared publications
Latest Publications
- 26 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Control Systems Letters
- Vol. 10, 13-18
Distributionally Robust GMM Steering Under Wasserstein Ambiguity Sets
This letter addresses the problem of steering a discrete-time linear system from an initial Gaussian Mixture Model (GMM) distribution, where the component weights are uncertain, to a prescribed terminal Gaussian distribution. We formulate a distributionally robust optimal control (DRO) problem by modeling the weight uncertainty via a Wasserstein-type ambiguity set. The resulting non-convex min-max problem is shown to be equivalently transformed into a tractable, deterministic Semidefinite Program (SDP) by leveraging duality theory and convex lifting techniques. The standard nominal (non-robust) controller can fail to satisfy terminal constraints under plausible worst-case weight distributions within the ambiguity set. In contrast, the proposed DRO controller successfully guarantees constraint satisfaction in these scenarios, demonstrating its improved robustness.
- 26 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Control Systems Letters
- Vol. 10, 13-18
Distributionally Robust GMM Steering Under Wasserstein Ambiguity Sets
This letter addresses the problem of steering a discrete-time linear system from an initial Gaussian Mixture Model (GMM) distribution, where the component weights are uncertain, to a prescribed terminal Gaussian distribution. We formulate a distributionally robust optimal control (DRO) problem by modeling the weight uncertainty via a Wasserstein-type ambiguity set. The resulting non-convex min-max problem is shown to be equivalently transformed into a tractable, deterministic Semidefinite Program (SDP) by leveraging duality theory and convex lifting techniques. The standard nominal (non-robust) controller can fail to satisfy terminal constraints under plausible worst-case weight distributions within the ambiguity set. In contrast, the proposed DRO controller successfully guarantees constraint satisfaction in these scenarios, demonstrating its improved robustness.
- 12 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (5) , 11834-11848
Robust Nonfragile Negative Imaginary H∞ Synthesis for Attitude Stabilization of Flexible Spacecrafts With Input Constraint
High-precision attitude stabilization of flexible spacecraft with model uncertainty, external disturbance, actuator fault and input constraint is investigated in this paper. A robust non-fragile negative imaginary (NI) $H_\infty$ synthesis scheme is developed, rendering the closed-loop system asymptotically stable and NI with $H_\infty$ norm bounded. Meanwhile, the control input $u$ is constrained by the specified upper bound. With a convex relaxation of Bilinear Matrix Inequalities (BMIs), the controller synthesis is cast as a convex optimization problem subject to Linear Matrix Inequality (LMI) constraints. The notable feature of the proposed method is that it achieves flexible vibration suppression and high-precision attitude stabilization simultaneously without requiring additional measurements or intelligent materials. The convergence of the LMIs-based iterative algorithm is guaranteed by a rigorous proof. Finally, comparative simulations demonstrate the effectiveness of the proposed method with high steady-state accuracy and low energy consumption.
- 12 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (5) , 11834-11848
Robust Nonfragile Negative Imaginary H∞ Synthesis for Attitude Stabilization of Flexible Spacecrafts With Input Constraint
High-precision attitude stabilization of flexible spacecraft with model uncertainty, external disturbance, actuator fault and input constraint is investigated in this paper. A robust non-fragile negative imaginary (NI) $H_\infty$ synthesis scheme is developed, rendering the closed-loop system asymptotically stable and NI with $H_\infty$ norm bounded. Meanwhile, the control input $u$ is constrained by the specified upper bound. With a convex relaxation of Bilinear Matrix Inequalities (BMIs), the controller synthesis is cast as a convex optimization problem subject to Linear Matrix Inequality (LMI) constraints. The notable feature of the proposed method is that it achieves flexible vibration suppression and high-precision attitude stabilization simultaneously without requiring additional measurements or intelligent materials. The convergence of the LMIs-based iterative algorithm is guaranteed by a rigorous proof. Finally, comparative simulations demonstrate the effectiveness of the proposed method with high steady-state accuracy and low energy consumption.
- Open Access
Pursuit-Interception Strategy in Differential Games Based on Q-Learning-Cover Algorithm
Due to the limited difference in maneuverability between the pursuer and the evader in three-dimensional space, it is difficult for a single pursuer to capture the evader. To address this, this paper proposes a strategy where three pursuers intercept one evader and introduces a Q-learning-cover algorithm. Based on the motion models of the pursuers and the evader in three-dimensional space, this paper presents a region coverage scheme based on the Ahlswede ball and analyzes the convergence upper bound of the Q-learning-cover algorithm by designing an appropriate Lyapunov function. Through extensive model training, the successful capture of the evader by the pursuers in a three-on-one scenario was achieved. Finally, numerical simulation experiments and hardware-in-the-loop simulation experiments are presented, both of which demonstrate that the proposed Q-learning-cover algorithm can effectively realize the three-on-one encirclement and interception of the evading target.
Open AccessPursuit-Interception Strategy in Differential Games Based on Q-Learning-Cover Algorithm
Due to the limited difference in maneuverability between the pursuer and the evader in three-dimensional space, it is difficult for a single pursuer to capture the evader. To address this, this paper proposes a strategy where three pursuers intercept one evader and introduces a Q-learning-cover algorithm. Based on the motion models of the pursuers and the evader in three-dimensional space, this paper presents a region coverage scheme based on the Ahlswede ball and analyzes the convergence upper bound of the Q-learning-cover algorithm by designing an appropriate Lyapunov function. Through extensive model training, the successful capture of the evader by the pursuers in a three-on-one scenario was achieved. Finally, numerical simulation experiments and hardware-in-the-loop simulation experiments are presented, both of which demonstrate that the proposed Q-learning-cover algorithm can effectively realize the three-on-one encirclement and interception of the evading target.
- Open Access
Optimal Pursuit Strategies in Missile Interception: Mean Field Game Approach
This paper investigates Mean Field Game methods to solve missile interception strategies in three-dimensional space, with a focus on analyzing the pursuit–evasion problem in many-to-many scenarios. By extending traditional missile interception models, an efficient solution is proposed to avoid dimensional explosion and communication burdens, particularly for large-scale, multi-missile systems. The paper presents a system of stochastic differential equations with control constraints, describing the motion dynamics between the missile (pursuer) and the target (evader), and defines the associated cost function, considering proximity group distributions with other missiles and targets. Next, Hamilton–Jacobi–Bellman equations for the pursuers and evaders are derived, and the uniqueness of the distributional solution is proved. Furthermore, using the -Nash equilibrium framework, it is demonstrated that, under the MFG model, participants can deviate from the optimal strategy within a certain tolerance, while still minimizing the cost. Finally, the paper summarizes the derivation process of the optimal strategy and proves that, under reasonable assumptions, the system can achieve a uniquely stable equilibrium, ensuring the stability of the strategies and distributions of both the pursuers and evaders. The research provides a scalable solution to high-risk, multi-agent control problems, with significant practical applications, particularly in fields such as missile defense systems.Open AccessOptimal Pursuit Strategies in Missile Interception: Mean Field Game Approach
This paper investigates Mean Field Game methods to solve missile interception strategies in three-dimensional space, with a focus on analyzing the pursuit–evasion problem in many-to-many scenarios. By extending traditional missile interception models, an efficient solution is proposed to avoid dimensional explosion and communication burdens, particularly for large-scale, multi-missile systems. The paper presents a system of stochastic differential equations with control constraints, describing the motion dynamics between the missile (pursuer) and the target (evader), and defines the associated cost function, considering proximity group distributions with other missiles and targets. Next, Hamilton–Jacobi–Bellman equations for the pursuers and evaders are derived, and the uniqueness of the distributional solution is proved. Furthermore, using the -Nash equilibrium framework, it is demonstrated that, under the MFG model, participants can deviate from the optimal strategy within a certain tolerance, while still minimizing the cost. Finally, the paper summarizes the derivation process of the optimal strategy and proves that, under reasonable assumptions, the system can achieve a uniquely stable equilibrium, ensuring the stability of the strategies and distributions of both the pursuers and evaders. The research provides a scalable solution to high-risk, multi-agent control problems, with significant practical applications, particularly in fields such as missile defense systems. - 18 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-11
TEVIO: Thermal-Aided Event-Based Visual–Inertial Odometry for Robust State Estimation in Challenging Environments
Event-based visual odometry excels in high dynamic range scenarios but struggles in extremely low-light or low-contrast conditions, motivating the integration of thermal imaging. This paper presents TEVIO, a multi-modal system that fuses thermal imaging, event-based vision, and inertial measurements to address the challenges of visual-inertial odometry in low-light, high-dynamic-range, and low-texture environments. An enhanced time surface map (ETSM) improves feature extraction for high-motion and low-texture scenes. A parallel frequency-varied tracking framework then estimates the pose stably and in high precision. Extensive tests on public event camera datasets and real-world outdoor vehicle experiments show TEVIO’s superior tracking accuracy and robustness compared to state-of-the-art monocular methods like EVIO, enabling reliable pose estimation in conditions where conventional approaches fail. A video demonstration is available at https://youtu.be/RfWYU15WwsU.
- 18 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-11
TEVIO: Thermal-Aided Event-Based Visual–Inertial Odometry for Robust State Estimation in Challenging Environments
Event-based visual odometry excels in high dynamic range scenarios but struggles in extremely low-light or low-contrast conditions, motivating the integration of thermal imaging. This paper presents TEVIO, a multi-modal system that fuses thermal imaging, event-based vision, and inertial measurements to address the challenges of visual-inertial odometry in low-light, high-dynamic-range, and low-texture environments. An enhanced time surface map (ETSM) improves feature extraction for high-motion and low-texture scenes. A parallel frequency-varied tracking framework then estimates the pose stably and in high precision. Extensive tests on public event camera datasets and real-world outdoor vehicle experiments show TEVIO’s superior tracking accuracy and robustness compared to state-of-the-art monocular methods like EVIO, enabling reliable pose estimation in conditions where conventional approaches fail. A video demonstration is available at https://youtu.be/RfWYU15WwsU.
- 1 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Journal of Systems Engineering and Electronics
- Vol. 36 (1) , 233-255
Two-to-One Differential Game via Improved MOGWO
When the maneuverability of a pursuer is not significantly higher than that of an evader, it will be difficult to intercept the evader with only one pursuer. Therefore, this article adopts a two-to-one differential game strategy, the game of kind is generally considered to be angle-optimized, which allows unlimited turns, but these practices do not take into account the effect of acceleration, which does not correspond to the actual situation, thus, based on the angle-optimized, the acceleration optimization and the acceleration upper bound constraint are added into the game for consideration. A two-to-one differential game problem is proposed in the three-dimensional space, and an improved multi-objective grey wolf optimization (IMOGWO) algorithm is proposed to solve the optimal game point of this problem. With the equations that describe the relative motions between the pursuers and the evader in the three-dimensional space, a multi-objective function with constraints is given as the performance index to design an optimal strategy for the differential game. Then the optimal game point is solved by using the IMOGWO algorithm. It is proved based on Markov chains that with the IMOGWO, the Pareto solution set is the solution of the differential game. Finally, it is verified through simulations that the pursuers can capture the escapee, and via comparative experiments, it is shown that the IMOGWO algorithm performs well in terms of running time and memory usage.
- 1 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Journal of Systems Engineering and Electronics
- Vol. 36 (1) , 233-255
Two-to-One Differential Game via Improved MOGWO
When the maneuverability of a pursuer is not significantly higher than that of an evader, it will be difficult to intercept the evader with only one pursuer. Therefore, this article adopts a two-to-one differential game strategy, the game of kind is generally considered to be angle-optimized, which allows unlimited turns, but these practices do not take into account the effect of acceleration, which does not correspond to the actual situation, thus, based on the angle-optimized, the acceleration optimization and the acceleration upper bound constraint are added into the game for consideration. A two-to-one differential game problem is proposed in the three-dimensional space, and an improved multi-objective grey wolf optimization (IMOGWO) algorithm is proposed to solve the optimal game point of this problem. With the equations that describe the relative motions between the pursuers and the evader in the three-dimensional space, a multi-objective function with constraints is given as the performance index to design an optimal strategy for the differential game. Then the optimal game point is solved by using the IMOGWO algorithm. It is proved based on Markov chains that with the IMOGWO, the Pareto solution set is the solution of the differential game. Finally, it is verified through simulations that the pursuers can capture the escapee, and via comparative experiments, it is shown that the IMOGWO algorithm performs well in terms of running time and memory usage.
- Open Access
A Class of Pursuit Problems in 3D Space via Noncooperative Stochastic Differential Games
This paper investigates three-dimensional pursuit problems in noncooperative stochastic differential games. By introducing a novel polynomial value function capable of addressing high-dimensional dynamic systems, the forward–backward stochastic differential equations (FBSDEs) for optimal strategies are derived. The uniqueness of the value function under bounded control inputs is rigorously established as a theoretical foundation. The proposed methodology constructs optimal closed-loop feedback strategies for both pursuers and evaders, ensuring state convergence and solution uniqueness. Furthermore, the Lebesgue measure of the barrier surface is computed, enabling the design of strategies for scenarios involving multiple pursuers and evaders. To validate its applicability, the method is applied to missile interception games. Simulations confirm that the optimal strategies enable pursuers to consistently intercept evaders under stochastic dynamics, demonstrating the robustness and practical relevance of the approach in pursuit–evasion problems.
Open AccessA Class of Pursuit Problems in 3D Space via Noncooperative Stochastic Differential Games
This paper investigates three-dimensional pursuit problems in noncooperative stochastic differential games. By introducing a novel polynomial value function capable of addressing high-dimensional dynamic systems, the forward–backward stochastic differential equations (FBSDEs) for optimal strategies are derived. The uniqueness of the value function under bounded control inputs is rigorously established as a theoretical foundation. The proposed methodology constructs optimal closed-loop feedback strategies for both pursuers and evaders, ensuring state convergence and solution uniqueness. Furthermore, the Lebesgue measure of the barrier surface is computed, enabling the design of strategies for scenarios involving multiple pursuers and evaders. To validate its applicability, the method is applied to missile interception games. Simulations confirm that the optimal strategies enable pursuers to consistently intercept evaders under stochastic dynamics, demonstrating the robustness and practical relevance of the approach in pursuit–evasion problems.
- Open Access
Advancing Human Motion Recognition with SkeletonCLIP++: Weighted Video Feature Integration and Enhanced Contrastive Sample Discrimination
This paper introduces ‘SkeletonCLIP++’, an extension of our prior work in human action recognition, emphasizing the use of semantic information beyond traditional label-based methods. The innovation, ‘Weighted Frame Integration’ (WFI), shifts video feature computation from averaging to a weighted frame approach, enabling a more nuanced representation of human movements in line with semantic relevance. Another key development, ‘Contrastive Sample Identification’ (CSI), introduces a novel discriminative task within the model. This task involves identifying the most similar negative sample among positive ones, enhancing the model’s ability to distinguish between closely related actions. Incorporating the ‘BERT Text Encoder Integration’ (BTEI) leverages the pre-trained BERT model as our text encoder to refine the model’s performance. Empirical evaluations on HMDB-51, UCF-101, and NTU RGB+D 60 datasets illustrate positive improvements, especially in smaller datasets. ‘SkeletonCLIP++’ thus offers a refined approach to human action recognition, ensuring semantic integrity and detailed differentiation in video data analysis.
Open AccessAdvancing Human Motion Recognition with SkeletonCLIP++: Weighted Video Feature Integration and Enhanced Contrastive Sample Discrimination
This paper introduces ‘SkeletonCLIP++’, an extension of our prior work in human action recognition, emphasizing the use of semantic information beyond traditional label-based methods. The innovation, ‘Weighted Frame Integration’ (WFI), shifts video feature computation from averaging to a weighted frame approach, enabling a more nuanced representation of human movements in line with semantic relevance. Another key development, ‘Contrastive Sample Identification’ (CSI), introduces a novel discriminative task within the model. This task involves identifying the most similar negative sample among positive ones, enhancing the model’s ability to distinguish between closely related actions. Incorporating the ‘BERT Text Encoder Integration’ (BTEI) leverages the pre-trained BERT model as our text encoder to refine the model’s performance. Empirical evaluations on HMDB-51, UCF-101, and NTU RGB+D 60 datasets illustrate positive improvements, especially in smaller datasets. ‘SkeletonCLIP++’ thus offers a refined approach to human action recognition, ensuring semantic integrity and detailed differentiation in video data analysis.
- Open Access
Globally Composite-Learning-Based Intelligent Fast Finite-Time Control for Uncertain Strict-Feedback Systems with Nonlinearly Periodic Disturbances
This brief aims at the issue of globally composite-learning-based neural fast finite-time (F-FnT) tracking control for a class of uncertain systems in strict-feedback form subject to nonlinearly periodic disturbances. First, uncertain dynamics with periodic parameters are identified by incorporating Fourier series expansion (FSE) into an intelligent estimator, which leverages the feedback of newly designed prediction errors in updating weights to boost learning performance. Then, a novel switching mechanism is constructed to fulfill smooth switching from the composite FSE-based neural controller to robust control law when the inputs of the intelligent estimator transcend the valid approximation domain. By fusing the switching mechanism with an improved F-FnT backstepping algorithm, the globally F-FnT boundedness of all variables in the closed-loop system is guaranteed. Finally, a simulation study is conducted to evince the availability of the theoretical result.
Open AccessGlobally Composite-Learning-Based Intelligent Fast Finite-Time Control for Uncertain Strict-Feedback Systems with Nonlinearly Periodic Disturbances
This brief aims at the issue of globally composite-learning-based neural fast finite-time (F-FnT) tracking control for a class of uncertain systems in strict-feedback form subject to nonlinearly periodic disturbances. First, uncertain dynamics with periodic parameters are identified by incorporating Fourier series expansion (FSE) into an intelligent estimator, which leverages the feedback of newly designed prediction errors in updating weights to boost learning performance. Then, a novel switching mechanism is constructed to fulfill smooth switching from the composite FSE-based neural controller to robust control law when the inputs of the intelligent estimator transcend the valid approximation domain. By fusing the switching mechanism with an improved F-FnT backstepping algorithm, the globally F-FnT boundedness of all variables in the closed-loop system is guaranteed. Finally, a simulation study is conducted to evince the availability of the theoretical result.
- Open Access
Improving Small-Scale Human Action Recognition Performance Using a 3D Heatmap Volume
In recent years, skeleton-based human action recognition has garnered significant research attention, with proposed recognition or segmentation methods typically validated on large-scale coarse-grained action datasets. However, there remains a lack of research on the recognition of small-scale fine-grained human actions using deep learning methods, which have greater practical significance. To address this gap, we propose a novel approach based on heatmap-based pseudo videos and a unified, general model applicable to all modality datasets. Leveraging anthropometric kinematics as prior information, we extract common human motion features among datasets through an ad hoc pre-trained model. To overcome joint mismatch issues, we partition the human skeleton into five parts, a simple yet effective technique for information sharing. Our approach is evaluated on two datasets, including the public Nursing Activities and our self-built Tai Chi Action dataset. Results from linear evaluation protocol and fine-tuned evaluation demonstrate that our pre-trained model effectively captures common motion features among human actions and achieves steady and precise accuracy across all training settings, while mitigating network overfitting. Notably, our model outperforms state-of-the-art models in recognition accuracy when fusing joint and limb modality features along the channel dimension.
Open AccessImproving Small-Scale Human Action Recognition Performance Using a 3D Heatmap Volume
In recent years, skeleton-based human action recognition has garnered significant research attention, with proposed recognition or segmentation methods typically validated on large-scale coarse-grained action datasets. However, there remains a lack of research on the recognition of small-scale fine-grained human actions using deep learning methods, which have greater practical significance. To address this gap, we propose a novel approach based on heatmap-based pseudo videos and a unified, general model applicable to all modality datasets. Leveraging anthropometric kinematics as prior information, we extract common human motion features among datasets through an ad hoc pre-trained model. To overcome joint mismatch issues, we partition the human skeleton into five parts, a simple yet effective technique for information sharing. Our approach is evaluated on two datasets, including the public Nursing Activities and our self-built Tai Chi Action dataset. Results from linear evaluation protocol and fine-tuned evaluation demonstrate that our pre-trained model effectively captures common motion features among human actions and achieves steady and precise accuracy across all training settings, while mitigating network overfitting. Notably, our model outperforms state-of-the-art models in recognition accuracy when fusing joint and limb modality features along the channel dimension.