He Shen
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
Research Interests
He Shen has a total of 45 co-authors. Here are the top co-authors by number of shared publications:
- Hongcheng Lian
Northwestern Polytechnical University
9 shared publications - Jun Luo
Northwestern Polytechnical University
9 shared publications - Lehua Qi
School of Mechanical Engineering, Northwestern Polytechnical University, Xi'an, China
9 shared publications Ni Li
3 shared publications- Yixin Yang
School of Marine Science and Technology, Northwestern Polytechnical University, Xi’an 710072, China
3 shared publications - Ni Li
School of Aeronautics, Northwestern Polytechnical University, Xi'an 710072, China; National Key Laboratory of Aircraft Configuration Design, Xi'an 710072, China
2 shared publications
Latest Publications
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 51 (4) , 3710-3720
MaRIne: An Underwater Monocular Real-Time Object Detection and Geometric Estimation Framework Based on Motion Inversion
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 51 (4) , 3710-3720
MaRIne: An Underwater Monocular Real-Time Object Detection and Geometric Estimation Framework Based on Motion Inversion
- 8 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (8) , 9088-9095
Collective Intelligent Reinforcement Learning for Biomimetic Clapping Trajectory Optimization
The foreflippers of sea lions enable highly efficient and agile locomotion through a unique clapping propulsion mechanism. Nonetheless, identifying the optimal kinematics for these movements constitutes a significant challenge, owing to the highly nonlinear and dynamic fluid-structure interactions. To address this, we present a collective intelligent reinforcement learning method for generating clapping trajectories of a bionic foreflipper. The approach begins by modeling biological motions using periodic triaxial motions based on closed Bézier curves, enabling the representation of essential clapping dynamics with a small set of parameters. Subsequently, we propose a collective intelligent reinforcement learning algorithm to efficiently optimize the parameterized clapping trajectory. A comparative evaluation against conventional optimization algorithms reveals accelerated convergence and superior performance. The optimal trajectories are validated against biological motion data, confirming their physiological plausibility. Simulation results show that a significant 75% improvement in thrust impulse and a 70% reduction in convergence time compared to the baseline reinforcement learning method.
- 8 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (8) , 9088-9095
Collective Intelligent Reinforcement Learning for Biomimetic Clapping Trajectory Optimization
The foreflippers of sea lions enable highly efficient and agile locomotion through a unique clapping propulsion mechanism. Nonetheless, identifying the optimal kinematics for these movements constitutes a significant challenge, owing to the highly nonlinear and dynamic fluid-structure interactions. To address this, we present a collective intelligent reinforcement learning method for generating clapping trajectories of a bionic foreflipper. The approach begins by modeling biological motions using periodic triaxial motions based on closed Bézier curves, enabling the representation of essential clapping dynamics with a small set of parameters. Subsequently, we propose a collective intelligent reinforcement learning algorithm to efficiently optimize the parameterized clapping trajectory. A comparative evaluation against conventional optimization algorithms reveals accelerated convergence and superior performance. The optimal trajectories are validated against biological motion data, confirming their physiological plausibility. Simulation results show that a significant 75% improvement in thrust impulse and a 70% reduction in convergence time compared to the baseline reinforcement learning method.
- 24 June 2025
- Published by Elsevier in Ocean Engineering
- Vol. 338
Zero-shot lightweight submarine cable detection in side-scan sonar images
- 24 June 2025
- Published by Elsevier in Ocean Engineering
- Vol. 338
Zero-shot lightweight submarine cable detection in side-scan sonar images
- 15 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 50 (3) , 2345-2354
Rotation Invariant Sonar Image Segmentation for Undersea Cables
Undersea cable detection is a prerequisite for cable maintenance and repair. However, extracting cables from side-scan sonar images is challenging due to the lack of details and interference from seabed sediments. In this article, an automatic rotation-invariant segmentation method for undersea cables is proposed. First, a filter based on the curvelet transform is designed to extract features of cables automatically. Second, a 2-D constant false alarm rate detector is used for feature denoising. Third, a morphology repair method is proposed to fulfill features that have been missed during feature extraction and image denoising. Finally, the maximum connected area in images is retained for cable segmentation. Results show that the proposed method can extract cables accurately. Four performance indicators, including structural similarity index, precision, pixel accuracy, and intersection over union reach 0.9810, 0.6108, 0.8348, and 0.8915, respectively. Consistent performance has been observed in images with different cable postures.
- 15 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 50 (3) , 2345-2354
Rotation Invariant Sonar Image Segmentation for Undersea Cables
Undersea cable detection is a prerequisite for cable maintenance and repair. However, extracting cables from side-scan sonar images is challenging due to the lack of details and interference from seabed sediments. In this article, an automatic rotation-invariant segmentation method for undersea cables is proposed. First, a filter based on the curvelet transform is designed to extract features of cables automatically. Second, a 2-D constant false alarm rate detector is used for feature denoising. Third, a morphology repair method is proposed to fulfill features that have been missed during feature extraction and image denoising. Finally, the maximum connected area in images is retained for cable segmentation. Results show that the proposed method can extract cables accurately. Four performance indicators, including structural similarity index, precision, pixel accuracy, and intersection over union reach 0.9810, 0.6108, 0.8348, and 0.8915, respectively. Consistent performance has been observed in images with different cable postures.
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-13
Data-Driven Hydrodynamic Force Estimation of Clapping Propulsion With Distributed Sensing
Sea lions use unique clapping propulsion for highly efficient and exceptional agile motion. A deep understanding of how hydrodynamic forces were generated over their foreflippers is the key to revealing the working mechanism of clapping propulsion. However, there have not been any effective and efficient methods to measure the complex interactions between solid flippers and the surrounding fluid. This paper proposes a data-driven hydrodynamic force estimation method using measurement from sparsely distributed pressure sensors. This method can quickly and accurately predict hydrodynamic forces without requiring the complex computation of coupled fluid-solid interactions. First, a four-channeled data structure, consisting of the pressure measurement and position/orientation of individual sensors, is created to model the state of a biomimetic flipper. Second, five data-driven convolutional neural networks with different structures are designed to model the relationship between the propulsion force and pressured measurements from the sparsely distributed sensors. Finally, simulations are carried out to verify the effectiveness of the proposed hydrodynamic force estimation method. The results show that a small number of sensors can estimate propulsion force with good accuracy and speed. The prediction performances of five neural networks with different architectures are compared. Increasing the depth of the model is conducive to improving the prediction accuracy, but continuing to increase the depth will produce overfitting. Inception’s strategy of using multi-sized convolution kernels outperforms others, providing a higher accuracy while requiring a lighter network architecture.
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-13
Data-Driven Hydrodynamic Force Estimation of Clapping Propulsion With Distributed Sensing
Sea lions use unique clapping propulsion for highly efficient and exceptional agile motion. A deep understanding of how hydrodynamic forces were generated over their foreflippers is the key to revealing the working mechanism of clapping propulsion. However, there have not been any effective and efficient methods to measure the complex interactions between solid flippers and the surrounding fluid. This paper proposes a data-driven hydrodynamic force estimation method using measurement from sparsely distributed pressure sensors. This method can quickly and accurately predict hydrodynamic forces without requiring the complex computation of coupled fluid-solid interactions. First, a four-channeled data structure, consisting of the pressure measurement and position/orientation of individual sensors, is created to model the state of a biomimetic flipper. Second, five data-driven convolutional neural networks with different structures are designed to model the relationship between the propulsion force and pressured measurements from the sparsely distributed sensors. Finally, simulations are carried out to verify the effectiveness of the proposed hydrodynamic force estimation method. The results show that a small number of sensors can estimate propulsion force with good accuracy and speed. The prediction performances of five neural networks with different architectures are compared. Increasing the depth of the model is conducive to improving the prediction accuracy, but continuing to increase the depth will produce overfitting. Inception’s strategy of using multi-sized convolution kernels outperforms others, providing a higher accuracy while requiring a lighter network architecture.
- 24 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 929-934
EMAR: Edge-Assisted Multi-User Mobile Augmented Reality
Multi-user Augmented Reality (AR) applications allow multiple users to interact within the same physical environment. However, existing multi-user systems lack the ability to recognize the complex physical environment. It is also challenging to achieve real-time multi-user AR under dynamic network conditions and different computing workloads. In this paper, we present EMAR, an edge-assisted multi-user mobile AR system. The system uses an edge cooperative optimization strategy to improve the accuracy of scene recognition, and adopts adaptive offloading and AR dynamic update strategy to ensure the real-time and effectiveness of AR recovery. We implement EMAR on mobile smartphones and an edge server. We verify its performance in different environments. The results show that EMAR achieves precision above 84.7% and recall above 92% for accurately identifying the same scene. Compared to the baseline method, our system is more robust in the dynamic environment, showing significant performance improvements.
- 24 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 929-934
EMAR: Edge-Assisted Multi-User Mobile Augmented Reality
Multi-user Augmented Reality (AR) applications allow multiple users to interact within the same physical environment. However, existing multi-user systems lack the ability to recognize the complex physical environment. It is also challenging to achieve real-time multi-user AR under dynamic network conditions and different computing workloads. In this paper, we present EMAR, an edge-assisted multi-user mobile AR system. The system uses an edge cooperative optimization strategy to improve the accuracy of scene recognition, and adopts adaptive offloading and AR dynamic update strategy to ensure the real-time and effectiveness of AR recovery. We implement EMAR on mobile smartphones and an edge server. We verify its performance in different environments. The results show that EMAR achieves precision above 84.7% and recall above 92% for accurately identifying the same scene. Compared to the baseline method, our system is more robust in the dynamic environment, showing significant performance improvements.
- 19 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ASME Transactions on Mechatronics
- Vol. 29 (1) , 52-63
A Compact Embedded Flight Parameter Detection System for Small Soaring UAVs
This article proposes a compact, lightweight, low-cost, and energy-efficient system to detect the air-relative flight parameters, which are essential for flight environment awareness of small soaring unmanned air vehicles (UAVs). Inspired by that some large birds with similar flight characteristics use distributed mechanoreceptors on or around follicles for airflow sensing and flight situation awareness, the proposed system consists of a group of micro piezoresistive barometric pressure sensors embedded in the skin of the wings for local total pressure measurements and a microcontroller for data acquisition and processing. A paired and symmetric sensor distribution design along with an iterative semiempirical robust regression model that is insensitive to outliers is proposed, allowing the system to simultaneously detect all three air-relative flight parameters with a desirable accuracy by using a limited number of sensors. General guidelines together with a comprehensive fluid simulation of a soaring UAV wing model are used to determine the ideal locations for sensor placement. This article further investigates the influence of the sensors’ number and placement on the performance of the proposed system. Finally, a soaring UAV wing with the proposed flight parameter detection system is fabricated, and wind-tunnel experiments are conducted to study and verify the effectiveness of the system.
- 19 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ASME Transactions on Mechatronics
- Vol. 29 (1) , 52-63
A Compact Embedded Flight Parameter Detection System for Small Soaring UAVs
This article proposes a compact, lightweight, low-cost, and energy-efficient system to detect the air-relative flight parameters, which are essential for flight environment awareness of small soaring unmanned air vehicles (UAVs). Inspired by that some large birds with similar flight characteristics use distributed mechanoreceptors on or around follicles for airflow sensing and flight situation awareness, the proposed system consists of a group of micro piezoresistive barometric pressure sensors embedded in the skin of the wings for local total pressure measurements and a microcontroller for data acquisition and processing. A paired and symmetric sensor distribution design along with an iterative semiempirical robust regression model that is insensitive to outliers is proposed, allowing the system to simultaneously detect all three air-relative flight parameters with a desirable accuracy by using a limited number of sensors. General guidelines together with a comprehensive fluid simulation of a soaring UAV wing model are used to determine the ideal locations for sensor placement. This article further investigates the influence of the sensors’ number and placement on the performance of the proposed system. Finally, a soaring UAV wing with the proposed flight parameter detection system is fabricated, and wind-tunnel experiments are conducted to study and verify the effectiveness of the system.
- 5 May 2023
- Published by Springer Nature
- p. 271-304
Microstructure Evolution and Interface Bonding of Uniform Aluminum Droplet Deposition Manufacturing
- 5 May 2023
- Published by Springer Nature
- p. 271-304
Microstructure Evolution and Interface Bonding of Uniform Aluminum Droplet Deposition Manufacturing