Fuyuan Feng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Fuyuan Feng has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Jian Gao
State Key Laboratory of Precision Electronic Manufacturing Technology and Equipment, School of Electromechanical Engineering, Guangdong University of Technology, Guangzhou 510006, China
2 shared publications - Lanyu Zhang
State Key Laboratory of Precision Electronic Manufacturing Technology and Equipment, School of Electromechanical Engineering, Guangdong University of Technology, Guangzhou 510006, China
2 shared publications - Yuheng Luo
State Key Laboratory of Precision Electronic Manufacturing Technology & Equipment, Guangdong University of Technology, Guangzhou, China
2 shared publications - Disai Chen
State Key Laboratory of Precision Electronic Manufacturing Technology and Equipment, Guangdong University of Technology, Guangzhou, China
1 shared publication Xin Chen
State Key Laboratory of Precision Electronic Manufacturing Technology and Equipment, Guangdong University of Technology, Guangzhou, China
1 shared publication- Xin Chen
State Key Laboratory of Precision Electronics Manufacturing Technology and Equipment, Guangdong University of Technology, Guangzhou, China
1 shared publication
Latest Publications
- 19 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ASME Transactions on Mechatronics
- Vol. 31 (1) , 1057-1068
Optimal Generalized Linear Bias Compensation Method and Implementation for Nongeometric Errors in Multiaxis Motion Mechanisms
Error compensation is an effective approach for improving the positioning accuracy of multiaxis motion mechanisms. Existing kinematics-based compensation method usually neglects some nongeometric factors in modeling process and affects the mechanism’s accuracy. Numerical fitting is a common method for nongeometric error compensation, but there still exists certain fitting error and degrades the error compensation effect. To achieve an optimal error compensation effectiveness, this article proposes a generalized linear bias compensation method for multiaxis mechanisms. In this method, we analyze the error formation mechanism of the numerical fitting method, and establish a generalized linear bias error model on basis of geometric error model. With the generalized model, two bias factor expressions corresponding to the existing direct and hybrid methods are presented and their fitting error differences and compensation effects are compared. To further improve the fitting accuracy, we define a new bias factor expression for the generalized model by expanding the training network with the geometric error compensated data. Through theoretical analysis, we deduce that the newly defined bias factor can represent the nongeometric errors closely, and the nongeometric errors can therefore be compensated accurately. The proposed compensation method is verified on a 3PPR&PR hybrid five-axis mechanism. Simulation and experimental results show that compared to the existing methods, the proposed optimal bias factor compensation method can effectively compensate for the residual nongeometric errors, and significantly enhance the motion and positioning accuracy of the multiaxis mechanism.
- 19 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ASME Transactions on Mechatronics
- Vol. 31 (1) , 1057-1068
Optimal Generalized Linear Bias Compensation Method and Implementation for Nongeometric Errors in Multiaxis Motion Mechanisms
Error compensation is an effective approach for improving the positioning accuracy of multiaxis motion mechanisms. Existing kinematics-based compensation method usually neglects some nongeometric factors in modeling process and affects the mechanism’s accuracy. Numerical fitting is a common method for nongeometric error compensation, but there still exists certain fitting error and degrades the error compensation effect. To achieve an optimal error compensation effectiveness, this article proposes a generalized linear bias compensation method for multiaxis mechanisms. In this method, we analyze the error formation mechanism of the numerical fitting method, and establish a generalized linear bias error model on basis of geometric error model. With the generalized model, two bias factor expressions corresponding to the existing direct and hybrid methods are presented and their fitting error differences and compensation effects are compared. To further improve the fitting accuracy, we define a new bias factor expression for the generalized model by expanding the training network with the geometric error compensated data. Through theoretical analysis, we deduce that the newly defined bias factor can represent the nongeometric errors closely, and the nongeometric errors can therefore be compensated accurately. The proposed compensation method is verified on a 3PPR&PR hybrid five-axis mechanism. Simulation and experimental results show that compared to the existing methods, the proposed optimal bias factor compensation method can effectively compensate for the residual nongeometric errors, and significantly enhance the motion and positioning accuracy of the multiaxis mechanism.
- 17 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 72 (9) , 9260-9270
An Innovative Pose Decoupling Calibration Method for Five-Axis Hybrid Mechanism Through Rotation Vector Equivalence
Geometric error calibration is one effective way to improve the absolute positioning accuracy of multidegree-of-freedom motion system. Due to differences in dimension, magnitudes, and measuring precision between position and orientation, the existing calibration methods may minimize the overall pose error but cannot guarantee that each degree-of-freedom error optimally converges in its own space. This makes it challenging for the current method to achieve a complete and precise error compensation for pose error. Addressing this issue, this article proposes an equivalent rotation vector (ERV)-based pose decoupling calibration method and thus transforms the rigid body transformation motion model (RTMM) into a pose-separation motion model (PSMM) to achieve an independent representation of orientation and position. With the PSMM, we can calibrate position and orientation separately without coupling interference and can thus make the position and orientation error converge in their own measurement spaces. The proposed method was experimentally validated on a five-axis hybrid mechanism. Compared to the current variable scaling method, our decoupling calibration method further reduces position and orientation errors by 66.88% and 51.99%, respectively. The proposed method can solve the nonconverging ill-identification problem of current methods, significantly improve the calibration accuracy of the end-effector pose, and achieve an optimal error compensation for the hybrid mechanism.
- 17 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 72 (9) , 9260-9270
An Innovative Pose Decoupling Calibration Method for Five-Axis Hybrid Mechanism Through Rotation Vector Equivalence
Geometric error calibration is one effective way to improve the absolute positioning accuracy of multidegree-of-freedom motion system. Due to differences in dimension, magnitudes, and measuring precision between position and orientation, the existing calibration methods may minimize the overall pose error but cannot guarantee that each degree-of-freedom error optimally converges in its own space. This makes it challenging for the current method to achieve a complete and precise error compensation for pose error. Addressing this issue, this article proposes an equivalent rotation vector (ERV)-based pose decoupling calibration method and thus transforms the rigid body transformation motion model (RTMM) into a pose-separation motion model (PSMM) to achieve an independent representation of orientation and position. With the PSMM, we can calibrate position and orientation separately without coupling interference and can thus make the position and orientation error converge in their own measurement spaces. The proposed method was experimentally validated on a five-axis hybrid mechanism. Compared to the current variable scaling method, our decoupling calibration method further reduces position and orientation errors by 66.88% and 51.99%, respectively. The proposed method can solve the nonconverging ill-identification problem of current methods, significantly improve the calibration accuracy of the end-effector pose, and achieve an optimal error compensation for the hybrid mechanism.