Yongfei Feng
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
Yongfei Feng has a total of 54 co-authors. Here are the top co-authors by number of shared publications:
- Hongbo Wang
Fudan University
4 shared publications - Jianye Niu
Hebei Provincial Key Laboratory of Parallel Robot and Mechatronic System, Yanshan University, Qinhuangdao, China
4 shared publications - Fangyan Dong
Faculty of Mechanical Engineering & Mechanics, Ningbo University, Ningbo 315211, China
3 shared publications - Liangda Wu
Faculty of Mechanical Engineering & Mechanics, Ningbo University, Ningbo 315211, China
3 shared publications - Luige Vladareanu
Robotics and Mechatronics Department, Institute of Solid Mechanics of the Romanian Academy, 010141 Bucharest, Romania
3 shared publications Hongbo Wang
Hebei Provincial Key Laboratory of Parallel Robot and Mechatronic System, Yanshan University, Qinhuangdao, China
2 shared publications
Latest Publications
- 1 October 2026
- Published by Springer Nature in Journal of Mechanical Science and Technology
- Vol. 40 (10) , 7487-7504
Design, analysis and experimental verification of rigid and flexible series-parallel trunk rehabilitation robot
- 1 October 2026
- Published by Springer Nature in Journal of Mechanical Science and Technology
- Vol. 40 (10) , 7487-7504
Design, analysis and experimental verification of rigid and flexible series-parallel trunk rehabilitation robot
- Open Access
sEMG-Based Motion Intention Recognition for Interactive Upper Limb Nursing Assistance
Surface electromyography (sEMG) enables non-invasive acquisition of neuromuscular activity and has shown strong potential for motion intention recognition in human–machine interaction. However, achieving reliable and real-time decoding remains critical for interactive upper-limb assistance. This study presents a structured sEMG-based framework for motion intention recognition in upper-limb assistance tasks, integrating multi-channel acquisition, standardized preprocessing, time-domain feature extraction, and supervised learning. sEMG signals from four representative motions were collected, and eight time-domain features were extracted from denoised and segmented signal windows. A compact feature subset was identified through systematic evaluation. Five classifiers were benchmarked under consistent validation conditions, with Random Forest achieving the best performance and further optimized via K-fold cross-validation. The proposed method achieved an average intra-subject accuracy of 95.23% across eight subjects and 95.72% in online interactive validation. These results demonstrate that time-domain feature fusion combined with ensemble learning provides robust and efficient motion discrimination, highlighting its potential for real-time assistive and rehabilitation applications.
Open AccesssEMG-Based Motion Intention Recognition for Interactive Upper Limb Nursing Assistance
Surface electromyography (sEMG) enables non-invasive acquisition of neuromuscular activity and has shown strong potential for motion intention recognition in human–machine interaction. However, achieving reliable and real-time decoding remains critical for interactive upper-limb assistance. This study presents a structured sEMG-based framework for motion intention recognition in upper-limb assistance tasks, integrating multi-channel acquisition, standardized preprocessing, time-domain feature extraction, and supervised learning. sEMG signals from four representative motions were collected, and eight time-domain features were extracted from denoised and segmented signal windows. A compact feature subset was identified through systematic evaluation. Five classifiers were benchmarked under consistent validation conditions, with Random Forest achieving the best performance and further optimized via K-fold cross-validation. The proposed method achieved an average intra-subject accuracy of 95.23% across eight subjects and 95.72% in online interactive validation. These results demonstrate that time-domain feature fusion combined with ensemble learning provides robust and efficient motion discrimination, highlighting its potential for real-time assistive and rehabilitation applications.
- 20 June 2025
- Published by Springer Nature
- p. 353-362
Modularized Complex Anthropomorphic Robot for Navigation on Uneven Terrain with High Stability
- 20 June 2025
- Published by Springer Nature
- p. 353-362
Modularized Complex Anthropomorphic Robot for Navigation on Uneven Terrain with High Stability
- 1 January 2025
- Published by Chinese Journal of Mechanical Engineering in Journal of Mechanical Engineering
- Vol. 61 (15) , 261-274
Joint Angle Estimation of Rehabilitation Robot-assisted Step Movement Based on Muscle Coordination Features
- 1 January 2025
- Published by Chinese Journal of Mechanical Engineering in Journal of Mechanical Engineering
- Vol. 61 (15) , 261-274
Joint Angle Estimation of Rehabilitation Robot-assisted Step Movement Based on Muscle Coordination Features
- Open Access
- 14 December 2024
- Published by MDPI AG in Technologies
- Vol. 12 (12) , 261
EEG-Based Mobile Robot Control Using Deep Learning and ROS Integration
Efficient BCIs (Brain-Computer Interfaces) harnessing EEG (Electroencephalography) have shown potential in controlling mobile robots, also presenting new possibilities for assistive technologies. This study explores the integration of advanced deep learning models—ASTGCN, EEGNetv4, and a combined CNN-LSTM architecture—with ROS (Robot Operating System) to control a two-wheeled mobile robot. The models were trained using a published EEG dataset, which includes signals from subjects performing thought-based tasks. Each model was evaluated based on its accuracy, F1-score, and latency. The CNN-LSTM architecture model exhibited the best performance on the cross-subject strategy with an accuracy of 88.5%, demonstrating significant potential for real-time applications. Integration with ROS was facilitated through a custom middleware, enabling seamless translation of neural commands into robot movements. The findings indicate that the CNN-LSTM model not only outperforms existing EEG-based systems in terms of accuracy but also underscores the practical feasibility of implementing such systems in real-world scenarios. Considering its efficacy, CNN-LSTM shows a great potential for assistive technology in the future. This research contributes to the development of a more intuitive and accessible robotic control system, potentially enhancing the quality of life for individuals with mobility impairments.
- 14 December 2024
- Published by MDPI AG in Technologies
- Vol. 12 (12) , 261
Open AccessEEG-Based Mobile Robot Control Using Deep Learning and ROS Integration
Efficient BCIs (Brain-Computer Interfaces) harnessing EEG (Electroencephalography) have shown potential in controlling mobile robots, also presenting new possibilities for assistive technologies. This study explores the integration of advanced deep learning models—ASTGCN, EEGNetv4, and a combined CNN-LSTM architecture—with ROS (Robot Operating System) to control a two-wheeled mobile robot. The models were trained using a published EEG dataset, which includes signals from subjects performing thought-based tasks. Each model was evaluated based on its accuracy, F1-score, and latency. The CNN-LSTM architecture model exhibited the best performance on the cross-subject strategy with an accuracy of 88.5%, demonstrating significant potential for real-time applications. Integration with ROS was facilitated through a custom middleware, enabling seamless translation of neural commands into robot movements. The findings indicate that the CNN-LSTM model not only outperforms existing EEG-based systems in terms of accuracy but also underscores the practical feasibility of implementing such systems in real-world scenarios. Considering its efficacy, CNN-LSTM shows a great potential for assistive technology in the future. This research contributes to the development of a more intuitive and accessible robotic control system, potentially enhancing the quality of life for individuals with mobility impairments.
- Open Access
A Linear Rehabilitative Motion Planning Method with a Multi-Posture Lower-Limb Rehabilitation Robot
In rehabilitation, physicians plan lower-limb exercises via linear guidance. Ensuring efficacy and safety, they design patient-specific paths, carefully plotting smooth trajectories to minimize jerks. Replicating their precision in robotics is a major challenge. This study introduces a linear rehabilitation motion planning method designed for physicians to use a multi-posture lower-limb rehabilitation robot, encompassing both path and trajectory planning. By subdividing the lower limb’s action space into four distinct training sections and classifying this space, we articulate the correlation between linear trajectories and key joint rehabilitation metrics. Building upon this foundation, a rehabilitative path generation system is developed, anchored in joint rehabilitation indicators. Subsequently, high-order polynomial curves are employed to mimic the smooth continuity of traditional rehabilitation trajectories and joint motions. Furthermore, trajectory planning is refined through the resolution of a constrained quadratic optimization problem, aiming to minimize the abrupt jerks in the trajectory. The optimized trajectories derived from our experiments are compared with randomly generated trajectories, demonstrating the suitability of trajectory optimization for real-time rehabilitation trajectory planning. Additionally, we compare trajectories generated based on the two groups of joint rehabilitation indicators, indicating that the proposed path generation system effectively assists clinicians in executing efficient and precise robot-assisted rehabilitation path planning.
Open AccessA Linear Rehabilitative Motion Planning Method with a Multi-Posture Lower-Limb Rehabilitation Robot
In rehabilitation, physicians plan lower-limb exercises via linear guidance. Ensuring efficacy and safety, they design patient-specific paths, carefully plotting smooth trajectories to minimize jerks. Replicating their precision in robotics is a major challenge. This study introduces a linear rehabilitation motion planning method designed for physicians to use a multi-posture lower-limb rehabilitation robot, encompassing both path and trajectory planning. By subdividing the lower limb’s action space into four distinct training sections and classifying this space, we articulate the correlation between linear trajectories and key joint rehabilitation metrics. Building upon this foundation, a rehabilitative path generation system is developed, anchored in joint rehabilitation indicators. Subsequently, high-order polynomial curves are employed to mimic the smooth continuity of traditional rehabilitation trajectories and joint motions. Furthermore, trajectory planning is refined through the resolution of a constrained quadratic optimization problem, aiming to minimize the abrupt jerks in the trajectory. The optimized trajectories derived from our experiments are compared with randomly generated trajectories, demonstrating the suitability of trajectory optimization for real-time rehabilitation trajectory planning. Additionally, we compare trajectories generated based on the two groups of joint rehabilitation indicators, indicating that the proposed path generation system effectively assists clinicians in executing efficient and precise robot-assisted rehabilitation path planning.
- 4 November 2024
- Published by Springer Nature in Journal of Mechanical Science and Technology
- Vol. 38 (11) , 6319-6333
Kinematics and performance analysis of a parallel 4-UPS/RRR trunk rehabilitation robot
- 4 November 2024
- Published by Springer Nature in Journal of Mechanical Science and Technology
- Vol. 38 (11) , 6319-6333
Kinematics and performance analysis of a parallel 4-UPS/RRR trunk rehabilitation robot
- Open Access
- 4 November 2024
- Published by MDPI AG in Bioengineering
- Vol. 11 (11) , 1113
An Assist-as-Needed Control Strategy Based on a Subjective Intention Decline Model
In the rehabilitation training process for stroke patients, the level of excitement in the patient’s physiological state has a positive impact on the efficacy of the training. In order to improve patients’ initiative during training and prevent dependence on assistive systems, this study proposes an assist-as-needed control strategy based on a subjective intention decline model. The strategy primarily consists of two modules: a subjective intention decline control module and a limb movement assessment module. The subjective intention decline module collects surface electromyography (sEMG) data during patient training and optimizes support vector machine (SVM) using quantum particle swarm optimization (QPSO) algorithms to establish a subjective intention decline model. The limb movement assessment module collects information such as interaction force and position error during training and proposes a method for evaluating the motion state of the affected limb. This model combines traditional impedance control with a method for assessing limb movement and subjective status, automatically adjusting the level of assistive force on the affected limb in real time to enhance its active participation in tasks. Finally, we performed two verification experiments to assess the patient’s initiative in participating in the training. The experimental results show that the proposed method effectively reduced the average assist force by 65.66% for the traditional impedance control training system and effectively the average assist force by 35.2% for the control training system using only the assist force module based on force position information. At the same time, the accuracy of the subjective intention attenuation module established in the experiment to identify the fatigue level of the subjects reached 93.41%. Therefore, the proposed method effectively improves the initiative of trainers and also prevents patients from relying on the assist-as-needed control training system.
- 4 November 2024
- Published by MDPI AG in Bioengineering
- Vol. 11 (11) , 1113
Open AccessAn Assist-as-Needed Control Strategy Based on a Subjective Intention Decline Model
In the rehabilitation training process for stroke patients, the level of excitement in the patient’s physiological state has a positive impact on the efficacy of the training. In order to improve patients’ initiative during training and prevent dependence on assistive systems, this study proposes an assist-as-needed control strategy based on a subjective intention decline model. The strategy primarily consists of two modules: a subjective intention decline control module and a limb movement assessment module. The subjective intention decline module collects surface electromyography (sEMG) data during patient training and optimizes support vector machine (SVM) using quantum particle swarm optimization (QPSO) algorithms to establish a subjective intention decline model. The limb movement assessment module collects information such as interaction force and position error during training and proposes a method for evaluating the motion state of the affected limb. This model combines traditional impedance control with a method for assessing limb movement and subjective status, automatically adjusting the level of assistive force on the affected limb in real time to enhance its active participation in tasks. Finally, we performed two verification experiments to assess the patient’s initiative in participating in the training. The experimental results show that the proposed method effectively reduced the average assist force by 65.66% for the traditional impedance control training system and effectively the average assist force by 35.2% for the control training system using only the assist force module based on force position information. At the same time, the accuracy of the subjective intention attenuation module established in the experiment to identify the fatigue level of the subjects reached 93.41%. Therefore, the proposed method effectively improves the initiative of trainers and also prevents patients from relying on the assist-as-needed control training system.
- Open Access
Advanced Intelligent Control in Robots
Advanced intelligent control (AIC) is a rapidly evolving and complex field that poses significant challenges
Open AccessAdvanced Intelligent Control in Robots
Advanced intelligent control (AIC) is a rapidly evolving and complex field that poses significant challenges
- 1 January 2023
- Published by ACTA Press in International Journal of Robotics and Automation
- Vol. 38 (2) , 85-96
KINEMATIC ANALYSIS OF THE VARIABLE HEIGHT ROTATING MECHANISM FOR END-TRACTION UPPER LIMB REHABILITATION ROBOT, 85-96.
To improve the motion diversity of planar end-traction rehabilitationrobot, a variable height rotating mechanism for the end-tractionupper limb rehabilitation robot (ULRR) is proposed based onthe serial–parallel mechanism.
- 1 January 2023
- Published by ACTA Press in International Journal of Robotics and Automation
- Vol. 38 (2) , 85-96
KINEMATIC ANALYSIS OF THE VARIABLE HEIGHT ROTATING MECHANISM FOR END-TRACTION UPPER LIMB REHABILITATION ROBOT, 85-96.
To improve the motion diversity of planar end-traction rehabilitationrobot, a variable height rotating mechanism for the end-tractionupper limb rehabilitation robot (ULRR) is proposed based onthe serial–parallel mechanism.