Yang Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Yang Liu has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Songlin Chen
Harbin Institute of Technology
7 shared publications - Cheng Xie
Control and Simulation Center, School of Astronautics, Harbin Institute of Technology, Harbin, China
6 shared publications Jiabao Geng
3 shared publications- Bing Xue
Centre for Data Science and Artificial Intelligence & School of Engineering and Computer Science, Victoria University of Wellington, New Zealand
2 shared publications - Mengjie Zhang
Centre for Data Science and Artificial Intelligence & School of Engineering and Computer Science, Victoria University of Wellington, New Zealand
2 shared publications Cheng Xie
Control and Simulation Center, School of Astronautics, Harbin Institute of Technology, Harbin, China
1 shared publication
Latest Publications
- 6 October 2026
- Published by Elsevier in Neurocomputing
- Vol. 709
Staleness-robust policy gradient via prefix correction and adaptive filtering in distributed reinforcement learning
- 6 October 2026
- Published by Elsevier in Neurocomputing
- Vol. 709
Staleness-robust policy gradient via prefix correction and adaptive filtering in distributed reinforcement learning
- 4 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (6) , 9957-9975
Enhancing Physical Human–Robot Interaction Compliance in Force Sensorless Lead-Through Programming for Direct Drive Manipulators
Lead-through programming (LTP) enables the operators to guide the manipulator directly to the desired target position, significantly improving adaptability and flexibility. However, direct drive manipulators encounter significant challenges in force sensorless LTP due to the high inertia, external force sensitivity, and complications from kinematics singularities. To enhance the interaction compliance and safety, a novel force sensorless LTP control strategy is proposed for direct drive manipulators. Specifically, an enhanced generalized momentum observer (EGMO) based on a nominal model is designed to mitigate phase lag caused by filtering measurement noise through a phase compensation mechanism. Additionally, the introduction of dither signal and radial basis function neural network (RBFNN) learning methods reduces dynamic uncertainties, improving control performance. Then, a kinematics singularity avoidance control method is proposed, which switches between impedance and admittance control by predicting singularity points in advance, enabling the manipulator to avoid kinematic singularity in the task space with minimal position and orientation deviations, promoting a more compliant and safe interaction between the manipulator and the human operator. Finally, the proposed methods are validated on a six-degree-of-freedom (DOF) direct drive manipulator under three typical singular scenarios. The results demonstrate the effectiveness and practicality of the proposed method.
- 4 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (6) , 9957-9975
Enhancing Physical Human–Robot Interaction Compliance in Force Sensorless Lead-Through Programming for Direct Drive Manipulators
Lead-through programming (LTP) enables the operators to guide the manipulator directly to the desired target position, significantly improving adaptability and flexibility. However, direct drive manipulators encounter significant challenges in force sensorless LTP due to the high inertia, external force sensitivity, and complications from kinematics singularities. To enhance the interaction compliance and safety, a novel force sensorless LTP control strategy is proposed for direct drive manipulators. Specifically, an enhanced generalized momentum observer (EGMO) based on a nominal model is designed to mitigate phase lag caused by filtering measurement noise through a phase compensation mechanism. Additionally, the introduction of dither signal and radial basis function neural network (RBFNN) learning methods reduces dynamic uncertainties, improving control performance. Then, a kinematics singularity avoidance control method is proposed, which switches between impedance and admittance control by predicting singularity points in advance, enabling the manipulator to avoid kinematic singularity in the task space with minimal position and orientation deviations, promoting a more compliant and safe interaction between the manipulator and the human operator. Finally, the proposed methods are validated on a six-degree-of-freedom (DOF) direct drive manipulator under three typical singular scenarios. The results demonstrate the effectiveness and practicality of the proposed method.
- 27 January 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 172
Evolving temporal convolutional neural networks for modeling turntable servo systems
- 27 January 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 172
Evolving temporal convolutional neural networks for modeling turntable servo systems
- 27 November 2024
- Published by SAGE Publications in Transactions of the Institute of Measurement and Control
- Vol. 48 (3) , 469-481
A feedforward control method for turntable servo system based on physics-guided neural network
Model-based feedforward control is a critical control method for high-dynamic and high-precision requirements of turntable servo systems. However, the turntable servo system will be affected by the change of working state in actual work, so the traditional physical feedforward control makes it difficult to obtain an accurate model of the turntable servo system. In this article, a feedforward controller based on a physics-guided neural network (PGNN) is proposed to improve the tracking performance of the turntable servo system. Firstly, a PGNN model is constructed according to the characteristics of the turntable, which is divided into a physical model part and a neural network part. Then the above model is applied to the design of the feedforward controller to replace the traditional feedforward controller. In order to improve the extension of the network, a new penalty function mechanism is proposed, which integrates regularization and physical consistency on the basis of the mean square error term. The automatic tuning algorithm of the parameters of the PGNN model is also designed to facilitate its wide application in the field of turntable control. Simulation results show that the neural network model designed in this article can accurately describe the nonlinear characteristics of the turntable servo system, and the fitting effect under different forms of excitation signals is better than the traditional physical model, with strong generalization ability and robustness. In addition, the feedforward control based on the model effectively improves the tracking accuracy of the system, which has important value and significance for the intelligent development of traditional servo control systems.
- 27 November 2024
- Published by SAGE Publications in Transactions of the Institute of Measurement and Control
- Vol. 48 (3) , 469-481
A feedforward control method for turntable servo system based on physics-guided neural network
Model-based feedforward control is a critical control method for high-dynamic and high-precision requirements of turntable servo systems. However, the turntable servo system will be affected by the change of working state in actual work, so the traditional physical feedforward control makes it difficult to obtain an accurate model of the turntable servo system. In this article, a feedforward controller based on a physics-guided neural network (PGNN) is proposed to improve the tracking performance of the turntable servo system. Firstly, a PGNN model is constructed according to the characteristics of the turntable, which is divided into a physical model part and a neural network part. Then the above model is applied to the design of the feedforward controller to replace the traditional feedforward controller. In order to improve the extension of the network, a new penalty function mechanism is proposed, which integrates regularization and physical consistency on the basis of the mean square error term. The automatic tuning algorithm of the parameters of the PGNN model is also designed to facilitate its wide application in the field of turntable control. Simulation results show that the neural network model designed in this article can accurately describe the nonlinear characteristics of the turntable servo system, and the fitting effect under different forms of excitation signals is better than the traditional physical model, with strong generalization ability and robustness. In addition, the feedforward control based on the model effectively improves the tracking accuracy of the system, which has important value and significance for the intelligent development of traditional servo control systems.
- 7 January 2024
- Published by Wiley in International Journal of Adaptive Control and Signal Processing
- Vol. 38 (3) , 1098-1113
Parameters self‐turning controller fused with an adaptive nominal model for disturbance observer: Application to direct drive manipulator with significant payload changes
Summary: The direct drive manipulator, which directly links the joint shaft to the motor rotor, is susceptible to external disturbances and model perturbations. When the direct drive manipulator operates with varying payloads, the disturbance observer (DOB) with a constant nominal model struggles to achieve satisfactory performance. To minimize estimation errors caused by disturbance, enhance trajectory tracking accuracy, and improve system robustness, this article proposes a controller parameter self‐tuning strategy with settling time, fused with a nominal model adaptive disturbance observer. Specifically, by optimizing the gain coefficient of the model reference adaptive identification (MRAI), the significant changes in payload at the manipulator's end joint can be quickly and accurately estimated. The inertia introduced by the payload is then employed to continuously update the nominal model of the manipulator, incorporating information about the inertia of the coupling joints. In addition, the controller parameters are updated using a self‐tuning link with settling time, which guarantees that the closed‐loop characteristics of the system remain unaffected by payload changes. Finally, the proposed method was validated using a two‐joint direct drive manipulator. The results demonstrate the robustness of the proposed method in the presence of significant payload changes, and the accuracy of disturbance estimation and the tracking performance of the proposed control strategy is significantly improved.
- 7 January 2024
- Published by Wiley in International Journal of Adaptive Control and Signal Processing
- Vol. 38 (3) , 1098-1113
Parameters self‐turning controller fused with an adaptive nominal model for disturbance observer: Application to direct drive manipulator with significant payload changes
Summary: The direct drive manipulator, which directly links the joint shaft to the motor rotor, is susceptible to external disturbances and model perturbations. When the direct drive manipulator operates with varying payloads, the disturbance observer (DOB) with a constant nominal model struggles to achieve satisfactory performance. To minimize estimation errors caused by disturbance, enhance trajectory tracking accuracy, and improve system robustness, this article proposes a controller parameter self‐tuning strategy with settling time, fused with a nominal model adaptive disturbance observer. Specifically, by optimizing the gain coefficient of the model reference adaptive identification (MRAI), the significant changes in payload at the manipulator's end joint can be quickly and accurately estimated. The inertia introduced by the payload is then employed to continuously update the nominal model of the manipulator, incorporating information about the inertia of the coupling joints. In addition, the controller parameters are updated using a self‐tuning link with settling time, which guarantees that the closed‐loop characteristics of the system remain unaffected by payload changes. Finally, the proposed method was validated using a two‐joint direct drive manipulator. The results demonstrate the robustness of the proposed method in the presence of significant payload changes, and the accuracy of disturbance estimation and the tracking performance of the proposed control strategy is significantly improved.
- 5 May 2023
- Published by Springer Nature in International Journal of Control, Automation, and Systems
- Vol. 21 (7) , 2363-2377
Modeling of Friction Torque Based on Turntable Servo System Using Neural Network with Memory Storage Unit
- 5 May 2023
- Published by Springer Nature in International Journal of Control, Automation, and Systems
- Vol. 21 (7) , 2363-2377
Modeling of Friction Torque Based on Turntable Servo System Using Neural Network with Memory Storage Unit