Zhiyang Liu
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
Zhiyang Liu has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Hongye Su
Zhejiang University
3 shared publications - Jian Chen
Southern University of Science and Technology
3 shared publications - Lei Wang
Zhejiang University
3 shared publications - Haisong Xu
College of Control Science and Engineering, Zhejiang University, Hangzhou, China
2 shared publications - Hao Chen
University of Shanghai for Science and Technology,School of Mechanical Engineering,Shanghai,China
2 shared publications Adn Marcelo H. Ang.
The Advanced Robotics Centre National Univer- sity of Singapore 117608 Singapore
1 shared publication
Latest Publications
- 4 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 12 (3) , 4388-4396
A Data-Driven Prediagnosis Framework for Early Detection of Water Imbalance in PEMFC
This article presents a prediagnosis approach for identifying water imbalance faults in proton exchange membrane fuel-cell (PEMFC) systems, aiming to provide early warning alarms to support proactive protection measures. Unlike traditional fixed-model methods, the proposed approach introduces a dynamic framework to handle the complexity of PEMFC fault conditions with improved efficiency and accuracy. First, the minimal redundancy maximal relevance (mRMR) algorithm is employed for feature selection to optimally balance input data quality and quantity. Subsequently, multiple encoder layers enhanced by a mixture of experts (MoE) mechanism are utilized to extract key information. By dynamically activating only the most relevant experts, the model’s adaptability to diverse input is improved, and computational load is reduced. Next, an adaptive multitask learning (MTL) framework is developed to train time-series forecasting and fault classification tasks simultaneously. Finally, the proposed method is validated on a 5-kW PEMFC stack under various operating conditions. Experimental results demonstrate that the proposed method achieves rapid and accurate identification of diverse PEMFC water management faults, detecting anomalies 25 s before occurrence with approximately 98.0% diagnostic accuracy at a fast response time of 1.15 ms.
- 4 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 12 (3) , 4388-4396
A Data-Driven Prediagnosis Framework for Early Detection of Water Imbalance in PEMFC
This article presents a prediagnosis approach for identifying water imbalance faults in proton exchange membrane fuel-cell (PEMFC) systems, aiming to provide early warning alarms to support proactive protection measures. Unlike traditional fixed-model methods, the proposed approach introduces a dynamic framework to handle the complexity of PEMFC fault conditions with improved efficiency and accuracy. First, the minimal redundancy maximal relevance (mRMR) algorithm is employed for feature selection to optimally balance input data quality and quantity. Subsequently, multiple encoder layers enhanced by a mixture of experts (MoE) mechanism are utilized to extract key information. By dynamically activating only the most relevant experts, the model’s adaptability to diverse input is improved, and computational load is reduced. Next, an adaptive multitask learning (MTL) framework is developed to train time-series forecasting and fault classification tasks simultaneously. Finally, the proposed method is validated on a 5-kW PEMFC stack under various operating conditions. Experimental results demonstrate that the proposed method achieves rapid and accurate identification of diverse PEMFC water management faults, detecting anomalies 25 s before occurrence with approximately 98.0% diagnostic accuracy at a fast response time of 1.15 ms.
- 1 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 72 (7) , 7544-7552
Online Identification of PEMFC Polarization Curves With Nonlinear Concentration Loss
Polarization curve plays a significant role in proton exchange membrane fuel cells (PEMFCs) monitoring and control. However, the classical semiempirical model of polarization curve is nonseparably parameterizated due to its nonlinear concentration loss. To transform the model into a linear regression equation (LRE), a differential operator-based method is introduced. Then, for the estimation of the derived LRE, four online estimators including gradient descent (GD) estimator, least squares (LS) estimator, dynamic regressor extension and mixing-based least squares (DREM+LS) estimator, and composite learning (CL) estimator are employed and compared in practical PEMFC systems. To further validate their universality, two experiments with different operating conditions were conducted on two different PEMFCs. Based on the experimental results, an in-depth comparative analysis of these estimators is provided to verify theoretical results. The experimental results reveal that, compared with the classical GD and LS estimators, DREM+LS and CL estimators achieved better performance in both steady and transient states, showing a smoother tracking process and faster response speed.
- 1 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 72 (7) , 7544-7552
Online Identification of PEMFC Polarization Curves With Nonlinear Concentration Loss
Polarization curve plays a significant role in proton exchange membrane fuel cells (PEMFCs) monitoring and control. However, the classical semiempirical model of polarization curve is nonseparably parameterizated due to its nonlinear concentration loss. To transform the model into a linear regression equation (LRE), a differential operator-based method is introduced. Then, for the estimation of the derived LRE, four online estimators including gradient descent (GD) estimator, least squares (LS) estimator, dynamic regressor extension and mixing-based least squares (DREM+LS) estimator, and composite learning (CL) estimator are employed and compared in practical PEMFC systems. To further validate their universality, two experiments with different operating conditions were conducted on two different PEMFCs. Based on the experimental results, an in-depth comparative analysis of these estimators is provided to verify theoretical results. The experimental results reveal that, compared with the classical GD and LS estimators, DREM+LS and CL estimators achieved better performance in both steady and transient states, showing a smoother tracking process and faster response speed.
- 23 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 72 (4) , 4053-4062
Fault Diagnosis for Proton Exchange Membrane Fuel Cell Systems via Improved Equivalent Circuit Model and Multiple Support Vector Machines
Fault diagnosis is of paramount importance in ensuring the reliability and stability of proton exchange membrane fuel cell systems. In this article, an innovative fault diagnosis method based on an improved equivalent circuit model (ECM) and multiple support vector machines (multi-SVMs) arranged in series is proposed. This method can enhance the interpretability of the ECM and feature data while achieving more efficient and accurate fault diagnosis. Initiating the process, the anode/cathode activation impedance, oxygen mass transfer impedance, and ohmic impedance are considered to design an improved ECM with a streamlined set of parameters for identification. The advantage of the improved ECM in terms of accuracy is demonstrated by comparison with two recently proposed ECMs. Then, a novel parameter identification method, integrating simulated annealing and complex nonlinear least square approach, is introduced to identify ECM parameters more efficiently and accurately. Finally, the multi-SVM classifier is built by concatenating multiple binary SVM models, then trained and validated using measured operating parameters from a 60 kW physical test bench and identified ECM parameters as feature data. Through a comparative evaluation with other classical classifiers, the advantages of multi-SVM in terms of training efficiency and classification accuracy can be demonstrated with training time of 0.3 s and test accuracy of 96.15%.
- 23 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 72 (4) , 4053-4062
Fault Diagnosis for Proton Exchange Membrane Fuel Cell Systems via Improved Equivalent Circuit Model and Multiple Support Vector Machines
Fault diagnosis is of paramount importance in ensuring the reliability and stability of proton exchange membrane fuel cell systems. In this article, an innovative fault diagnosis method based on an improved equivalent circuit model (ECM) and multiple support vector machines (multi-SVMs) arranged in series is proposed. This method can enhance the interpretability of the ECM and feature data while achieving more efficient and accurate fault diagnosis. Initiating the process, the anode/cathode activation impedance, oxygen mass transfer impedance, and ohmic impedance are considered to design an improved ECM with a streamlined set of parameters for identification. The advantage of the improved ECM in terms of accuracy is demonstrated by comparison with two recently proposed ECMs. Then, a novel parameter identification method, integrating simulated annealing and complex nonlinear least square approach, is introduced to identify ECM parameters more efficiently and accurately. Finally, the multi-SVM classifier is built by concatenating multiple binary SVM models, then trained and validated using measured operating parameters from a 60 kW physical test bench and identified ECM parameters as feature data. Through a comparative evaluation with other classical classifiers, the advantages of multi-SVM in terms of training efficiency and classification accuracy can be demonstrated with training time of 0.3 s and test accuracy of 96.15%.
- Open Access
You Only Scan Once: A Dynamic Scene Reconstruction Pipeline for 6-DoF Robotic Grasping of Novel Objects
In the realm of robotic grasping, achieving accurate and reliable interactions with the environment is a pivotal challenge. Traditional methods of grasp planning methods utilizing partial point clouds derived from depth image often suffer from reduced scene understanding due to occlusion, ultimately impeding their grasping accuracy. Furthermore, scene reconstruction methods have primarily relied upon static techniques, which are susceptible to environment change during manipulation process limits their efficacy in real-time grasping tasks. To address these limitations, this paper introduces a novel two-stage pipeline for dynamic scene reconstruction. In the first stage, our approach takes scene scanning as input to register each target object with mesh reconstruction and novel object pose tracking. In the second stage, pose tracking is still performed to provide object poses in real-time, enabling our approach to transform the reconstructed object point clouds back into the scene. Unlike conventional methodologies, which rely on static scene snapshots, our method continuously captures the evolving scene geometry, resulting in a comprehensive and up-to-date point cloud representation. By circumventing the constraints posed by occlusion, our method enhances the overall grasp planning process and empowers state-of-the-art 6-DoF robotic grasping algorithms to exhibit markedly improved accuracy.
Open AccessYou Only Scan Once: A Dynamic Scene Reconstruction Pipeline for 6-DoF Robotic Grasping of Novel Objects
In the realm of robotic grasping, achieving accurate and reliable interactions with the environment is a pivotal challenge. Traditional methods of grasp planning methods utilizing partial point clouds derived from depth image often suffer from reduced scene understanding due to occlusion, ultimately impeding their grasping accuracy. Furthermore, scene reconstruction methods have primarily relied upon static techniques, which are susceptible to environment change during manipulation process limits their efficacy in real-time grasping tasks. To address these limitations, this paper introduces a novel two-stage pipeline for dynamic scene reconstruction. In the first stage, our approach takes scene scanning as input to register each target object with mesh reconstruction and novel object pose tracking. In the second stage, pose tracking is still performed to provide object poses in real-time, enabling our approach to transform the reconstructed object point clouds back into the scene. Unlike conventional methodologies, which rely on static scene snapshots, our method continuously captures the evolving scene geometry, resulting in a comprehensive and up-to-date point cloud representation. By circumventing the constraints posed by occlusion, our method enhances the overall grasp planning process and empowers state-of-the-art 6-DoF robotic grasping algorithms to exhibit markedly improved accuracy.
- Open Access
A Scalable Segmented-Based PEM Fuel Cell Hybrid Power System Model and Its Simulation Applications
A scalable segmented-based proton-exchange membrane fuel cell (PEMFC) hybrid power system model is developed in this paper. The fuel cell (FC) is developed as a dynamic lumped parameter model to predict the current distributions during dynamic load scenarios. The fuel cell is segmented into segments connected with several physical ports and with the variables balanced automatically. Based on the proposed model, a real-time energy management framework is designed to distribute the load current demand during dynamic operations. Simulation results show that the proposed strategy has good performance on both single/segmented fuel cell–battery hybrid systems and the low battery state of charge (SOC) situation. This paper proposes an approach that uses an interconnected ordinary differential equations (ODEs) system model in control problems, which makes the control algorithms readily applicable.Open AccessA Scalable Segmented-Based PEM Fuel Cell Hybrid Power System Model and Its Simulation Applications
A scalable segmented-based proton-exchange membrane fuel cell (PEMFC) hybrid power system model is developed in this paper. The fuel cell (FC) is developed as a dynamic lumped parameter model to predict the current distributions during dynamic load scenarios. The fuel cell is segmented into segments connected with several physical ports and with the variables balanced automatically. Based on the proposed model, a real-time energy management framework is designed to distribute the load current demand during dynamic operations. Simulation results show that the proposed strategy has good performance on both single/segmented fuel cell–battery hybrid systems and the low battery state of charge (SOC) situation. This paper proposes an approach that uses an interconnected ordinary differential equations (ODEs) system model in control problems, which makes the control algorithms readily applicable. - Open Access
GET-DIPP: Graph-Embedded Transformer for Differentiable Integrated Prediction and Planning
Accurately predicting interactive road agents' future trajectories and planning a socially compliant and human-like trajectory accordingly are important for autonomous vehicles. In this paper, we propose a planning-centric prediction neural network, which takes surrounding agents' historical states and map context information as input, and outputs the joint multi-modal prediction trajectories for surrounding agents, as well as a sequence of control commands for the ego vehicle by imitation learning. An agent-agent interaction module along the time axis is proposed in our network architecture to better comprehend the relationship among all the other intelligent agents on the road. To incorporate the map's topological information, a Dynamic Graph Convolutional Neural Network (DGCNN) is employed to process the road network topology. Besides, the whole architecture can serve as a backbone for the Differentiable Integrated motion Prediction with Planning (DIPP) method by providing accurate prediction results and initial planning commands. Experiments are conducted on real-world datasets to demonstrate the improvements made by our proposed method in both planning and prediction accuracy compared to the previous state-of-the-art methods.
Open AccessGET-DIPP: Graph-Embedded Transformer for Differentiable Integrated Prediction and Planning
Accurately predicting interactive road agents' future trajectories and planning a socially compliant and human-like trajectory accordingly are important for autonomous vehicles. In this paper, we propose a planning-centric prediction neural network, which takes surrounding agents' historical states and map context information as input, and outputs the joint multi-modal prediction trajectories for surrounding agents, as well as a sequence of control commands for the ego vehicle by imitation learning. An agent-agent interaction module along the time axis is proposed in our network architecture to better comprehend the relationship among all the other intelligent agents on the road. To incorporate the map's topological information, a Dynamic Graph Convolutional Neural Network (DGCNN) is employed to process the road network topology. Besides, the whole architecture can serve as a backbone for the Differentiable Integrated motion Prediction with Planning (DIPP) method by providing accurate prediction results and initial planning commands. Experiments are conducted on real-world datasets to demonstrate the improvements made by our proposed method in both planning and prediction accuracy compared to the previous state-of-the-art methods.
- 11 May 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ASME Transactions on Mechatronics
- Vol. 25 (4) , 2009-2018
A Modified MPC-Based Optimal Strategy of Power Management for Fuel Cell Hybrid Vehicles
Power management is a vital technology to increase the economy of fuel cell hybrid vehicles. To overcome the deficiency of most existing optimization approaches, a modified model predictive control-based optimal strategy of power management is proposed for fuel cell hybrid vehicles. Besides, a partial least squares-based load estimation method is utilized to avoid the assumption of conventional model predictive control. Furthermore, an alternating direction multiplier method-based quadratic programming solver is also designed to avoid the ill condition of the quadratic programming. Finally, the reliability and effectiveness of the proposed strategy are demonstrated by the simulation and experiments.
- 11 May 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ASME Transactions on Mechatronics
- Vol. 25 (4) , 2009-2018
A Modified MPC-Based Optimal Strategy of Power Management for Fuel Cell Hybrid Vehicles
Power management is a vital technology to increase the economy of fuel cell hybrid vehicles. To overcome the deficiency of most existing optimization approaches, a modified model predictive control-based optimal strategy of power management is proposed for fuel cell hybrid vehicles. Besides, a partial least squares-based load estimation method is utilized to avoid the assumption of conventional model predictive control. Furthermore, an alternating direction multiplier method-based quadratic programming solver is also designed to avoid the ill condition of the quadratic programming. Finally, the reliability and effectiveness of the proposed strategy are demonstrated by the simulation and experiments.
- 25 March 2019
- Published by Wiley in Asian Journal of Control
- Vol. 21 (4) , 1847-1856
Real‐time optimal energy management for a fuel cell/battery hybrid system
An energy management strategy is proposed for a class of fuel cell/battery hybrid systems. In such hybrid systems, a fuel cell system is the main power source, and a lithium‐ion battery is the auxiliary power source. In order to manage the system power at the next moment in a reasonable way, a load current filter with bounded estimation errors is designed to estimate the load current. Then, a real‐time optimal energy management algorithm is proposed to optimize economy consumption of the hybrid system. By taking current change rate of the fuel cell and the state of charge into consideration and taking reasonable model simplifications, the optimization problem can be described as a quadratic programming problem. Then a general purpose solver is proposed to solve the quadratic programming problem based on the alternating direction method of multipliers. The efficiency of the proposed solver is much faster than computing interior point method or active set method. Simulation results in MATLAB/SIMULINK are carried out to validate the significant effectiveness and efficiency of the proposed management strategy.
- 25 March 2019
- Published by Wiley in Asian Journal of Control
- Vol. 21 (4) , 1847-1856
Real‐time optimal energy management for a fuel cell/battery hybrid system
An energy management strategy is proposed for a class of fuel cell/battery hybrid systems. In such hybrid systems, a fuel cell system is the main power source, and a lithium‐ion battery is the auxiliary power source. In order to manage the system power at the next moment in a reasonable way, a load current filter with bounded estimation errors is designed to estimate the load current. Then, a real‐time optimal energy management algorithm is proposed to optimize economy consumption of the hybrid system. By taking current change rate of the fuel cell and the state of charge into consideration and taking reasonable model simplifications, the optimization problem can be described as a quadratic programming problem. Then a general purpose solver is proposed to solve the quadratic programming problem based on the alternating direction method of multipliers. The efficiency of the proposed solver is much faster than computing interior point method or active set method. Simulation results in MATLAB/SIMULINK are carried out to validate the significant effectiveness and efficiency of the proposed management strategy.