Zhu Wang
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
Zhu Wang has a total of 39 co-authors. Here are the top co-authors by number of shared publications:
Li Liu
Key Laboratory of Dynamics and Control of Flight Vehicle, Ministry of Education, School of Aerospace Engineering, Beijing Institute of Technology, Beijing, People's Republic of China
9 shared publicationsTeng Long
School of Aerospace Engineering, Beijing Institute of Technology, Beijing 100081, P.R. China IEEE and Key Laboratory of Dynamics and Control of Flight Vehicle, Ministry of Education, Beijing 100081, P.R. China
9 shared publications- Guangtong Xu
Huzhou Institute, Zhejiang University, Huzhou, China
6 shared publications Jiaxun Kou
School of Aerospace Engineering, Beijing Institute of Technology 100081
5 shared publications- Teng Long
School of Aerospace Engineering, Beijing Institute of Technology, Beijing, China
5 shared publications Yonglu Wen
5 shared publications
Latest Publications
- Open Access
- 10 February 2026
- Published by Elsevier in Chinese Journal of Aeronautics
- Vol. 39 (8)
Gradient-based trajectory planner of real-time obstacle avoidance for fixed-wing unmanned aerial vehicles in cluttered environments
- 10 February 2026
- Published by Elsevier in Chinese Journal of Aeronautics
- Vol. 39 (8)
Open AccessGradient-based trajectory planner of real-time obstacle avoidance for fixed-wing unmanned aerial vehicles in cluttered environments
- 5 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. PP, 1-8
Adaptive-Risk-Aware End-to-End Control System for Mapless Navigation of Drones
Safe and autonomous navigation of drones in 3D complex unknown environments remains a current research hotspot. A critical challenge is how to adaptively balance the safety with flight aggressiveness, fully exploiting the motion potential of the drone. This letter proposes an Adaptive-Risk-Aware End-to-End (ARA-E2E) control system which can generate real-time control commands capable of achieving adaptive balance in 3D environments. The system utilizes the Deterministic Implicit Quantile Network (DIQN) algorithm under the Actor-Critic framework, directly generating acceleration commands from LiDAR data. In the critic network, the Implicit Quantile Networks (IQN) algorithm, which integrates a risk-aware module, is employed. IQN evaluates actions with varying tendencies toward flight speed and risk, by calculating the parameter of Conditional Value-at-Risk (CVaR) based on the current obstacle environment. In the actor network, a continuous state space representation method is designed to provide more flexible control commands. The comparative experiments show that the proposed method outperforms in flight time, success rate, and energy consumption in the forest scenario. To further validate the effectiveness of the system, we deploy it on an autonomous drone platform, and flight tests are conducted in an outdoor forest environment. The drone can successfully complete the real-world navigation task at an average speed of 2.5 m/s (with a maximum of 5 m/s in the simulation), demonstrating the practicality and reliability of the proposed method.
- 5 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. PP, 1-8
Adaptive-Risk-Aware End-to-End Control System for Mapless Navigation of Drones
Safe and autonomous navigation of drones in 3D complex unknown environments remains a current research hotspot. A critical challenge is how to adaptively balance the safety with flight aggressiveness, fully exploiting the motion potential of the drone. This letter proposes an Adaptive-Risk-Aware End-to-End (ARA-E2E) control system which can generate real-time control commands capable of achieving adaptive balance in 3D environments. The system utilizes the Deterministic Implicit Quantile Network (DIQN) algorithm under the Actor-Critic framework, directly generating acceleration commands from LiDAR data. In the critic network, the Implicit Quantile Networks (IQN) algorithm, which integrates a risk-aware module, is employed. IQN evaluates actions with varying tendencies toward flight speed and risk, by calculating the parameter of Conditional Value-at-Risk (CVaR) based on the current obstacle environment. In the actor network, a continuous state space representation method is designed to provide more flexible control commands. The comparative experiments show that the proposed method outperforms in flight time, success rate, and energy consumption in the forest scenario. To further validate the effectiveness of the system, we deploy it on an autonomous drone platform, and flight tests are conducted in an outdoor forest environment. The drone can successfully complete the real-world navigation task at an average speed of 2.5 m/s (with a maximum of 5 m/s in the simulation), demonstrating the practicality and reliability of the proposed method.
- Open Access
Flexible Combinatorial-Bids-Based Auction for Cooperative Target Assignment of Unmanned Aerial Vehicles
For the cooperative reconnaissance assignment of unmanned aerial vehicles (UAVs) on multiple targets, this paper presents a flexible combinatorial-bids-based auction (FCBA) method that can optimize the number of UAVs for each target. Considering the reconnaissance effectiveness enhancement achieved with cooperative observation and the time-critical characteristic of targets, the multitarget assignment problem is formulated as a nonlinear integer optimization to maximize the cooperative effectiveness. To achieve target assignment without predetermining the number of UAVs for each target, a combinatorial bidding framework is proposed, and an allocation method for rewards and costs among the cooperative UAVs is constructed. Strategies for auction iteration and bid updating are also designed to acquire equilibrium results under the combinatorial bidding mechanism. The simulation results show that the proposed method can generate satisfactory suboptimal results according to the enumerated solutions. A comparison of the results demonstrates that the FCBA can provide comparable optimal results to a genetic algorithm but has better computational efficiency, and the reconnaissance effectiveness can be improved by considering cooperative observation.
Open AccessFlexible Combinatorial-Bids-Based Auction for Cooperative Target Assignment of Unmanned Aerial Vehicles
For the cooperative reconnaissance assignment of unmanned aerial vehicles (UAVs) on multiple targets, this paper presents a flexible combinatorial-bids-based auction (FCBA) method that can optimize the number of UAVs for each target. Considering the reconnaissance effectiveness enhancement achieved with cooperative observation and the time-critical characteristic of targets, the multitarget assignment problem is formulated as a nonlinear integer optimization to maximize the cooperative effectiveness. To achieve target assignment without predetermining the number of UAVs for each target, a combinatorial bidding framework is proposed, and an allocation method for rewards and costs among the cooperative UAVs is constructed. Strategies for auction iteration and bid updating are also designed to acquire equilibrium results under the combinatorial bidding mechanism. The simulation results show that the proposed method can generate satisfactory suboptimal results according to the enumerated solutions. A comparison of the results demonstrates that the FCBA can provide comparable optimal results to a genetic algorithm but has better computational efficiency, and the reconnaissance effectiveness can be improved by considering cooperative observation.
- 17 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (2) , 1928-1935
Comparison of Distributed Task Allocation Algorithms Considering Non-Ideal Communication Factors for Multi-UAV Collaborative Visit Missions
This letter comprehensively investigates the performance of six state-of-art distributed task allocation algorithms (i.e., CBAA, CBBA, HIPC, PI, DHBA, and DGA) subject to non-ideal communication factors. The package loss, bit error, and time delay factors are considered in the distributed task allocation process. The performance of the algorithms for multi-UAV collaborative visit missions is compared under pre-allocation and dynamic allocation scenarios. The synchronous and asynchronous communication modes are separately utilized in different allocation scenarios for analyzing the effects of non-ideal communication factors. Comparison results show that bit error factors cause conflicted allocations. For the pre-allocation scenario, CBBA outperforms the competitors in terms of reliability, communication overhead, and efficiency. For the dynamic scenario, CBBA performs best optimality, while DHBA exhibits better reliability and lower overhead in harsh communication conditions.
- 17 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (2) , 1928-1935
Comparison of Distributed Task Allocation Algorithms Considering Non-Ideal Communication Factors for Multi-UAV Collaborative Visit Missions
This letter comprehensively investigates the performance of six state-of-art distributed task allocation algorithms (i.e., CBAA, CBBA, HIPC, PI, DHBA, and DGA) subject to non-ideal communication factors. The package loss, bit error, and time delay factors are considered in the distributed task allocation process. The performance of the algorithms for multi-UAV collaborative visit missions is compared under pre-allocation and dynamic allocation scenarios. The synchronous and asynchronous communication modes are separately utilized in different allocation scenarios for analyzing the effects of non-ideal communication factors. Comparison results show that bit error factors cause conflicted allocations. For the pre-allocation scenario, CBBA outperforms the competitors in terms of reliability, communication overhead, and efficiency. For the dynamic scenario, CBBA performs best optimality, while DHBA exhibits better reliability and lower overhead in harsh communication conditions.
- 28 November 2022
- Published by Elsevier in ISA Transactions
- Vol. 136, 715-726
Real-time multi-quadrotor trajectory generation via distributed receding architecture and hierarchical planning in complex environments
- 28 November 2022
- Published by Elsevier in ISA Transactions
- Vol. 136, 715-726
Real-time multi-quadrotor trajectory generation via distributed receding architecture and hierarchical planning in complex environments
- Open Access
Generalization Differences between End-to-End and Neuro-Symbolic Vision-Language Reasoning Systems
For vision-and-language reasoning tasks, both fully connectionist, end-to-end methods and hybrid, neuro-symbolic methods have achieved high in-distribution performance. In which out-of-distribution settings does each paradigm excel? We investigate this question on both single-image and multi-image visual question-answering through four types of generalization tests: a novel segment-combine test for multi-image queries, contrast set, compositional generalization, and cross-benchmark transfer. Vision-and-language end-to-end trained systems exhibit sizeable performance drops across all these tests. Neuro-symbolic methods suffer even more on cross-benchmark transfer from GQA to VQA, but they show smaller accuracy drops on the other generalization tests and their performance quickly improves by few-shot training. Overall, our results demonstrate the complementary benefits of these two paradigms, and emphasize the importance of using a diverse suite of generalization tests to fully characterize model robustness to distribution shift.
Open AccessGeneralization Differences between End-to-End and Neuro-Symbolic Vision-Language Reasoning Systems
For vision-and-language reasoning tasks, both fully connectionist, end-to-end methods and hybrid, neuro-symbolic methods have achieved high in-distribution performance. In which out-of-distribution settings does each paradigm excel? We investigate this question on both single-image and multi-image visual question-answering through four types of generalization tests: a novel segment-combine test for multi-image queries, contrast set, compositional generalization, and cross-benchmark transfer. Vision-and-language end-to-end trained systems exhibit sizeable performance drops across all these tests. Neuro-symbolic methods suffer even more on cross-benchmark transfer from GQA to VQA, but they show smaller accuracy drops on the other generalization tests and their performance quickly improves by few-shot training. Overall, our results demonstrate the complementary benefits of these two paradigms, and emphasize the importance of using a diverse suite of generalization tests to fully characterize model robustness to distribution shift.
- 1 September 2022
- Published by Elsevier in ISA Transactions
- Vol. 128 (Pt B) , 664-676
Trust-region filtered sequential convex programming for multi-UAV trajectory planning and collision avoidance
- 1 September 2022
- Published by Elsevier in ISA Transactions
- Vol. 128 (Pt B) , 664-676
Trust-region filtered sequential convex programming for multi-UAV trajectory planning and collision avoidance
- 22 March 2021
- Published by SPIE-Intl Soc Optical Eng
- p. 73
Experimental study on a novel corrosion sensor for steel strand using FBG
SPIE Digital Library Proceedings
- 22 March 2021
- Published by SPIE-Intl Soc Optical Eng
- p. 73
Experimental study on a novel corrosion sensor for steel strand using FBG
SPIE Digital Library Proceedings
- Open Access
- 1 November 2020
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1684 (1) , 012123
Image compression-encryption algorithm based on 2D compressive sensing and double random phase encoding
In this paper, an image compression and encryption algorithm based on 2D compressive sensing and double random phase encoding is proposed. Compared with traditional encryption algorithms, 2D compressive sensing can realize image compression and encryption at the same time, with high execution efficiency. Double random phase encoding can not only realize image encryption fast, but also has robustness. Moreover, combining 2D compressive sensing with double random phase encoding can obtain better encryption security. Simulation results show that this algorithm has better image compression performance, and can ensure security and robustness.
- 1 November 2020
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1684 (1) , 012123
Open AccessImage compression-encryption algorithm based on 2D compressive sensing and double random phase encoding
In this paper, an image compression and encryption algorithm based on 2D compressive sensing and double random phase encoding is proposed. Compared with traditional encryption algorithms, 2D compressive sensing can realize image compression and encryption at the same time, with high execution efficiency. Double random phase encoding can not only realize image encryption fast, but also has robustness. Moreover, combining 2D compressive sensing with double random phase encoding can obtain better encryption security. Simulation results show that this algorithm has better image compression performance, and can ensure security and robustness.
- Open Access
- 4 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 90941-90953
Matrix Structure Driven Interior Point Method for Quadrotor Real-Time Trajectory Planning
Sequential convex programming (SCP) has been recently employed to various trajectory planning problems including entry flight, planetary landing, and aircraft formation. In SCP, convex programming subproblems are sequentially solved to obtain optimum of original nonconvex problems. For SCP-based quadrotor trajectory planning, this paper proposes a matrix-structure driven interior point method (MSD-IPM) to improve the efficiency of solving search directions in convex programming. In MSD-IPM, primal-dual systems for solving search directions are derived from Karush-Kuhn-Tucher (KKT) conditions of quadrotor trajectory planning subproblems. Then, the successive elimination technique is used to solve the inverse of large-scale coefficient matrices of primal-dual systems by more efficient operations on small-scale matrices. In successive elimination, the positive definiteness of several small-scale matrices is used to enhance the numerical stability of computing search directions. And the specific diagonal structures of small-scale matrices are exploited to efficiently compute search directions. The complexity analysis shows that the efficiency of the proposed method is about one order of magnitude higher than that of the standard interior point method. The comparative studies on simulation experiments demonstrate that MSD-IPM generally outperforms several well-known optimizers (e.g., MOSEK, SDPT3, and SeDuMi) in terms of efficiency and robustness. Finally, indoor trajectory tracking experiments indicate the proposed method can generate smooth trajectories for real-world applications.
- 4 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 90941-90953
Open AccessMatrix Structure Driven Interior Point Method for Quadrotor Real-Time Trajectory Planning
Sequential convex programming (SCP) has been recently employed to various trajectory planning problems including entry flight, planetary landing, and aircraft formation. In SCP, convex programming subproblems are sequentially solved to obtain optimum of original nonconvex problems. For SCP-based quadrotor trajectory planning, this paper proposes a matrix-structure driven interior point method (MSD-IPM) to improve the efficiency of solving search directions in convex programming. In MSD-IPM, primal-dual systems for solving search directions are derived from Karush-Kuhn-Tucher (KKT) conditions of quadrotor trajectory planning subproblems. Then, the successive elimination technique is used to solve the inverse of large-scale coefficient matrices of primal-dual systems by more efficient operations on small-scale matrices. In successive elimination, the positive definiteness of several small-scale matrices is used to enhance the numerical stability of computing search directions. And the specific diagonal structures of small-scale matrices are exploited to efficiently compute search directions. The complexity analysis shows that the efficiency of the proposed method is about one order of magnitude higher than that of the standard interior point method. The comparative studies on simulation experiments demonstrate that MSD-IPM generally outperforms several well-known optimizers (e.g., MOSEK, SDPT3, and SeDuMi) in terms of efficiency and robustness. Finally, indoor trajectory tracking experiments indicate the proposed method can generate smooth trajectories for real-world applications.