Wenping Jiang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Wenping Jiang has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
Lingyang Wang
Shanghai Institute of Technology, School of Electrical and Electronic Engineering, No. 100 Haiquan Road, Fengxian District, Shanghai 201418, China
2 shared publications- Zhencun Jiang
Shanghai Institute of Technology, 100 Haiquan Road, Shanghai 201418, China
2 shared publications - Jian Sang
Shanghai Keanxun Optoelectronics Co., Ltd., Shanghai 200000, People’s Republic of China
1 shared publication Jin Liang
School of Electrical and Electronic Engineering, Shanghai Institute of Technology, Shanghai, China
1 shared publicationJun Min
Tongji University, No. 1239 Siping Road, Yangpu District, Shanghai 200082, China
1 shared publication- Jun Min
College of Electronic and Information Engineering, Tongji University, Shanghai, China
1 shared publication
Latest Publications
- 2 September 2026
- Published by Springer Nature
- p. 645-654
Overfitting Risks and Solutions in Deep Learning for Railway Fault Diagnosis
- 2 September 2026
- Published by Springer Nature
- p. 645-654
Overfitting Risks and Solutions in Deep Learning for Railway Fault Diagnosis
- Open Access
Data-Driven Method for Robotic Trajectory Error Prediction and Compensation Based on Digital Twin
In addressing the limited absolute positioning accuracy of industrial robots, which stems from the discrepancy between the nominal kinematic model and the physical entity, this paper proposes a novel paradigm for online compensation based on data-driven error prediction. The present study utilized a KUKA KR4 R600 robot as the experimental platform to construct a high-fidelity digital twin system capable of real-time synchronization. Within this framework, a new machine learning model, termed the Global Configuration-Error Forest (GCE-Forest), was developed and validated. The fundamental principle of GCE-Forest, based on the Random Forest algorithm, is its offline learning of the complex, highly non-linear mapping from the robot’s six-dimensional joint space configuration to its three-dimensional end-effector Cartesian error space. This facilitates online, feedforward, and predictive compensation for the nominal trajectory during robot operation. Through rigorous comparative experiments, the superiority of the proposed GCE-Forest was established. The final outcomes of dynamic trajectory tracking validation demonstrate that the system, by accurately predicting a mean nominal error of 0.1977 mm, successfully reduced the average spatial positioning error of the end-effector to 0.0845 mm, achieving an accuracy improvement of 57.25%. This research provides comprehensive validation of the method’s robust performance, offering a low-cost, non-invasive, and highly effective solution for significantly enhancing robotic accuracy.
Open AccessData-Driven Method for Robotic Trajectory Error Prediction and Compensation Based on Digital Twin
In addressing the limited absolute positioning accuracy of industrial robots, which stems from the discrepancy between the nominal kinematic model and the physical entity, this paper proposes a novel paradigm for online compensation based on data-driven error prediction. The present study utilized a KUKA KR4 R600 robot as the experimental platform to construct a high-fidelity digital twin system capable of real-time synchronization. Within this framework, a new machine learning model, termed the Global Configuration-Error Forest (GCE-Forest), was developed and validated. The fundamental principle of GCE-Forest, based on the Random Forest algorithm, is its offline learning of the complex, highly non-linear mapping from the robot’s six-dimensional joint space configuration to its three-dimensional end-effector Cartesian error space. This facilitates online, feedforward, and predictive compensation for the nominal trajectory during robot operation. Through rigorous comparative experiments, the superiority of the proposed GCE-Forest was established. The final outcomes of dynamic trajectory tracking validation demonstrate that the system, by accurately predicting a mean nominal error of 0.1977 mm, successfully reduced the average spatial positioning error of the end-effector to 0.0845 mm, achieving an accuracy improvement of 57.25%. This research provides comprehensive validation of the method’s robust performance, offering a low-cost, non-invasive, and highly effective solution for significantly enhancing robotic accuracy.
- 10 March 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (4)
A multi-window CUSUM method with adaptive threshold and variable-length window for nonintrusive load monitoring
Non-Intrusive Load Monitoring (NILM) disaggregates appliance-level energy consumption from aggregate load data, offering a practical approach for analyzing household energy use and promoting efficiency. Event detection, a critical element in NILM, identifies state transitions in appliances. However, existing approaches often struggle with real-world data, where challenges arise from subtle load changes, high-frequency noise, and load-switching transients, all of which impact detection accuracy. Additionally, high computational demands can limit real-time applicability and reduce adaptability to varied appliance behaviors. In this paper, we present a multi-window cumulative sum (CUSUM) method featuring adaptive thresholds and variable-length windows, enhanced by a dynamic parameter adjustment mechanism driven by signal statistics. To address noise interference, the method integrates median filtering and exponentially weighted moving average (EWMA) preprocessing. Dynamic parameter adjustments enable adaptation to fluctuating appliance behaviors, while multi-window accumulation and rate-of-change analysis improve discrimination between genuine events and spurious fluctuations. Experimental results demonstrate the proposed method's superiority over conventional CUSUM and other improved methods on PLAID, BLUED, and private datasets, achieving F1 scores of 94.72%, 96.43%, and 91.98%, respectively. Its robustness across thresholds, practicality, and low computational requirements highlight its potential for real-world NILM applications on resource-constrained devices. Keywords: NILM, event detection, multi-window, adaptive threshold, variable-length
- 10 March 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (4)
A multi-window CUSUM method with adaptive threshold and variable-length window for nonintrusive load monitoring
Non-Intrusive Load Monitoring (NILM) disaggregates appliance-level energy consumption from aggregate load data, offering a practical approach for analyzing household energy use and promoting efficiency. Event detection, a critical element in NILM, identifies state transitions in appliances. However, existing approaches often struggle with real-world data, where challenges arise from subtle load changes, high-frequency noise, and load-switching transients, all of which impact detection accuracy. Additionally, high computational demands can limit real-time applicability and reduce adaptability to varied appliance behaviors. In this paper, we present a multi-window cumulative sum (CUSUM) method featuring adaptive thresholds and variable-length windows, enhanced by a dynamic parameter adjustment mechanism driven by signal statistics. To address noise interference, the method integrates median filtering and exponentially weighted moving average (EWMA) preprocessing. Dynamic parameter adjustments enable adaptation to fluctuating appliance behaviors, while multi-window accumulation and rate-of-change analysis improve discrimination between genuine events and spurious fluctuations. Experimental results demonstrate the proposed method's superiority over conventional CUSUM and other improved methods on PLAID, BLUED, and private datasets, achieving F1 scores of 94.72%, 96.43%, and 91.98%, respectively. Its robustness across thresholds, practicality, and low computational requirements highlight its potential for real-world NILM applications on resource-constrained devices. Keywords: NILM, event detection, multi-window, adaptive threshold, variable-length
- Open Access
- 8 April 2024
- Published by Frontiers Media SA in Frontiers in Energy Research
- Vol. 12, 1367903
Control and physical verification of 6-DOF manipulator for power inspection robots based on expert PID algorithm
To enhance the performance of power inspection robots in intricate nuclear power stations, it is necessary to improve their response speed and accuracy. This paper uses the manipulator of the power inspection robot as the primary research object, and unlike previous control algorithm research, which only remained in the software simulation stage, we constructed a set of physical verification platforms based on CAN communication and physically verified the robotic arm’s control algorithm. First, the forward motion model is established based on the geometric structure of the manipulator and D-H parameter method, and the kinematic equation of the manipulator is solved by combining geometric method and algebraic method. Secondly, in order to conduct comparison tests, we designed PID controllers and expert PID controllers by utilising the expertise of experts. The results show that compared with the traditional PID algorithm, the expert PID algorithm has a faster response speed in the control process of the manipulator. It converges quickly in 0.75 s and has a smaller overshoot, with a maximum of only 6.9%. This confirms the expert PID algorithm’s good control effect on the robotic arm, allowing the six-degree-of-freedom robotic arm to travel more accurately and swiftly along the trajectory of the target point.
- 8 April 2024
- Published by Frontiers Media SA in Frontiers in Energy Research
- Vol. 12, 1367903
Open AccessControl and physical verification of 6-DOF manipulator for power inspection robots based on expert PID algorithm
To enhance the performance of power inspection robots in intricate nuclear power stations, it is necessary to improve their response speed and accuracy. This paper uses the manipulator of the power inspection robot as the primary research object, and unlike previous control algorithm research, which only remained in the software simulation stage, we constructed a set of physical verification platforms based on CAN communication and physically verified the robotic arm’s control algorithm. First, the forward motion model is established based on the geometric structure of the manipulator and D-H parameter method, and the kinematic equation of the manipulator is solved by combining geometric method and algebraic method. Secondly, in order to conduct comparison tests, we designed PID controllers and expert PID controllers by utilising the expertise of experts. The results show that compared with the traditional PID algorithm, the expert PID algorithm has a faster response speed in the control process of the manipulator. It converges quickly in 0.75 s and has a smaller overshoot, with a maximum of only 6.9%. This confirms the expert PID algorithm’s good control effect on the robotic arm, allowing the six-degree-of-freedom robotic arm to travel more accurately and swiftly along the trajectory of the target point.
- 2 January 2024
- Published by Springer Nature
- p. 31-48
New Approach for Soil Moisture Prediction Based on Multiple Influencing Factors
- 2 January 2024
- Published by Springer Nature
- p. 31-48
New Approach for Soil Moisture Prediction Based on Multiple Influencing Factors
- Open Access
- 16 October 2023
- Published by Frontiers Media SA in Frontiers in Plant Science
- Vol. 14, 1258658
A ResNet50-DPA model for tomato leaf disease identification
Tomato leaf disease identification is difficult owing to the variety of diseases and complex causes, for which the method based on the convolutional neural network is effective. While it is challenging to capture key features or tends to lose a large number of features when extracting image features by applying this method, resulting in low accuracy of disease identification. Therefore, the ResNet50-DPA model is proposed to identify tomato leaf diseases in the paper. Firstly, an improved ResNet50 is included in the model, which replaces the first layer of convolution in the basic ResNet50 model with the cascaded atrous convolution, facilitating to obtaining of leaf features with different scales. Secondly, in the model, a dual-path attention (DPA) mechanism is proposed to search for key features, where the stochastic pooling is employed to eliminate the influence of non-maximum values, and two convolutions with one dimension are introduced to replace the MLP layer for effectively reducing the damage to leaf information. In addition, to quickly and accurately identify the type of leaf disease, the DPA module is incorporated into the residual module of the improved ResNet50 to obtain an enhanced tomato leaf feature map, which helps to reduce economic losses. Finally, the visualization results of Grad-CAM are presented to show that the ResNet50-DPA model proposed can identify diseases more accurately and improve the interpretability of the model, meeting the need for precise identification of tomato leaf diseases.
- 16 October 2023
- Published by Frontiers Media SA in Frontiers in Plant Science
- Vol. 14, 1258658
Open AccessA ResNet50-DPA model for tomato leaf disease identification
Tomato leaf disease identification is difficult owing to the variety of diseases and complex causes, for which the method based on the convolutional neural network is effective. While it is challenging to capture key features or tends to lose a large number of features when extracting image features by applying this method, resulting in low accuracy of disease identification. Therefore, the ResNet50-DPA model is proposed to identify tomato leaf diseases in the paper. Firstly, an improved ResNet50 is included in the model, which replaces the first layer of convolution in the basic ResNet50 model with the cascaded atrous convolution, facilitating to obtaining of leaf features with different scales. Secondly, in the model, a dual-path attention (DPA) mechanism is proposed to search for key features, where the stochastic pooling is employed to eliminate the influence of non-maximum values, and two convolutions with one dimension are introduced to replace the MLP layer for effectively reducing the damage to leaf information. In addition, to quickly and accurately identify the type of leaf disease, the DPA module is incorporated into the residual module of the improved ResNet50 to obtain an enhanced tomato leaf feature map, which helps to reduce economic losses. Finally, the visualization results of Grad-CAM are presented to show that the ResNet50-DPA model proposed can identify diseases more accurately and improve the interpretability of the model, meeting the need for precise identification of tomato leaf diseases.
- Open Access
Linear Golden Section Speed Adaptive Control of Permanent Magnet Synchronous Motor Based on Model Design
Permanent magnet synchronous motor (PMSM) is a multi-variable, strongly coupled, nonlinear complex system. It is usually difficult to establish an accurate mathematical model, and the introduction of new complex algorithms will increase the difficulty of embedded code development. In order to solve this problem, we establish the characteristic model of permanent magnet synchronous motor in this paper, and the speed control scheme of the linear golden-section adaptive control and integral compensation, which is adopted. Finally, using the model-based design (MBD) method, how to build the simulink embedded code automatic generation model is introduced in detail, and then we complete the PMSM speed control physical verification experiment. Simulation and experimental results show that compared with traditional proportional-integral-derivative (PID) control, the speed control accuracy of PMSM is improved about 3.8 times. Meanwhile, the development method based on the model design can increase the PMSM control system physical verification, and then improve the development efficiency.
Open AccessLinear Golden Section Speed Adaptive Control of Permanent Magnet Synchronous Motor Based on Model Design
Permanent magnet synchronous motor (PMSM) is a multi-variable, strongly coupled, nonlinear complex system. It is usually difficult to establish an accurate mathematical model, and the introduction of new complex algorithms will increase the difficulty of embedded code development. In order to solve this problem, we establish the characteristic model of permanent magnet synchronous motor in this paper, and the speed control scheme of the linear golden-section adaptive control and integral compensation, which is adopted. Finally, using the model-based design (MBD) method, how to build the simulink embedded code automatic generation model is introduced in detail, and then we complete the PMSM speed control physical verification experiment. Simulation and experimental results show that compared with traditional proportional-integral-derivative (PID) control, the speed control accuracy of PMSM is improved about 3.8 times. Meanwhile, the development method based on the model design can increase the PMSM control system physical verification, and then improve the development efficiency.
- Open Access
- 19 August 2021
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2021, 1-9
Application of Characteristic Model-Based Principal Component Analysis in Optimization of Flowmeter Parameters
In complex industrial processes, it is necessary to perform modeling analysis on some industrial systems and find and optimize the factors that have the greatest impact on the results, in order to achieve the optimization of the industrial systems. However, due to the high-level nature or complex working mechanism of complex industrial systems, traditional principal component analysis methods are difficult to apply. Therefore, this paper proposes a characteristic model-based principal component analysis (CMPCA) to perform principal component analysis on complex industrial systems. The differential pressure flowmeter is taken as an example to verify the effectiveness of the method. Flowmeter is an indispensable instrument in measurement, and its accuracy depends on its own structural parameters. However, the measurement accuracy of some flow meters is not high, and the measurement error is large, which affects the normal industrial production process. This method is used to analyze the influence of the structural parameters of the flowmeter on its measurement accuracy, and the four most important structural parameters are found and optimized. The measurement error of the Bitoba flowmeter is reduced from 1% to 0.2%, and the measurement repeatability is reduced from 0.3 to 0.06, which proves the effectiveness of the method.
- 19 August 2021
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2021, 1-9
Open AccessApplication of Characteristic Model-Based Principal Component Analysis in Optimization of Flowmeter Parameters
In complex industrial processes, it is necessary to perform modeling analysis on some industrial systems and find and optimize the factors that have the greatest impact on the results, in order to achieve the optimization of the industrial systems. However, due to the high-level nature or complex working mechanism of complex industrial systems, traditional principal component analysis methods are difficult to apply. Therefore, this paper proposes a characteristic model-based principal component analysis (CMPCA) to perform principal component analysis on complex industrial systems. The differential pressure flowmeter is taken as an example to verify the effectiveness of the method. Flowmeter is an indispensable instrument in measurement, and its accuracy depends on its own structural parameters. However, the measurement accuracy of some flow meters is not high, and the measurement error is large, which affects the normal industrial production process. This method is used to analyze the influence of the structural parameters of the flowmeter on its measurement accuracy, and the four most important structural parameters are found and optimized. The measurement error of the Bitoba flowmeter is reduced from 1% to 0.2%, and the measurement repeatability is reduced from 0.3 to 0.06, which proves the effectiveness of the method.
- Open Access
- 1 January 2021
- Published by Wiley in Computational Intelligence and Neuroscience
- Vol. 2021 (1)
Method for Diagnosis of Acute Lymphoblastic Leukemia Based on ViT‐CNN Ensemble Model
Acute lymphocytic leukemia (ALL) is a deadly cancer that not only affects adults but also accounts for about 25% of childhood cancers. Timely and accurate diagnosis of the cancer is an important premise for effective treatment to improve survival rate. Since the image of leukemic B‐lymphoblast cells (cancer cells) under the microscope is very similar in morphology to that of normal B‐lymphoid precursors (normal cells), it is difficult to distinguish between cancer cells and normal cells. Therefore, we propose the ViT‐CNN ensemble model to classify cancer cells images and normal cells images to assist in the diagnosis of acute lymphoblastic leukemia. The ViT‐CNN ensemble model is an ensemble model that combines the vision transformer model and convolutional neural network (CNN) model. The vision transformer model is an image classification model based entirely on the transformer structure, which has completely different feature extraction method from the CNN model. The ViT‐CNN ensemble model can extract the features of cells images in two completely different ways to achieve better classification results. In addition, the data set used in this article is an unbalanced data set and has a certain amount of noise, and we propose a difference enhancement‐random sampling (DERS) data enhancement method, create a new balanced data set, and use the symmetric cross‐entropy loss function to reduce the impact of noise in the data set. The classification accuracy of the ViT‐CNN ensemble model on the test set has reached 99.03%, and it is proved through experimental comparison that the effect is better than other models. The proposed method can accurately distinguish between cancer cells and normal cells and can be used as an effective method for computer‐aided diagnosis of acute lymphoblastic leukemia.
- 1 January 2021
- Published by Wiley in Computational Intelligence and Neuroscience
- Vol. 2021 (1)
Open AccessMethod for Diagnosis of Acute Lymphoblastic Leukemia Based on ViT‐CNN Ensemble Model
Acute lymphocytic leukemia (ALL) is a deadly cancer that not only affects adults but also accounts for about 25% of childhood cancers. Timely and accurate diagnosis of the cancer is an important premise for effective treatment to improve survival rate. Since the image of leukemic B‐lymphoblast cells (cancer cells) under the microscope is very similar in morphology to that of normal B‐lymphoid precursors (normal cells), it is difficult to distinguish between cancer cells and normal cells. Therefore, we propose the ViT‐CNN ensemble model to classify cancer cells images and normal cells images to assist in the diagnosis of acute lymphoblastic leukemia. The ViT‐CNN ensemble model is an ensemble model that combines the vision transformer model and convolutional neural network (CNN) model. The vision transformer model is an image classification model based entirely on the transformer structure, which has completely different feature extraction method from the CNN model. The ViT‐CNN ensemble model can extract the features of cells images in two completely different ways to achieve better classification results. In addition, the data set used in this article is an unbalanced data set and has a certain amount of noise, and we propose a difference enhancement‐random sampling (DERS) data enhancement method, create a new balanced data set, and use the symmetric cross‐entropy loss function to reduce the impact of noise in the data set. The classification accuracy of the ViT‐CNN ensemble model on the test set has reached 99.03%, and it is proved through experimental comparison that the effect is better than other models. The proposed method can accurately distinguish between cancer cells and normal cells and can be used as an effective method for computer‐aided diagnosis of acute lymphoblastic leukemia.
- Open Access
- 2 June 2019
- Published by Wiley in Journal of Control Science and Engineering
- Vol. 2019, 1-8
Design of Winding Parameters Based on Multiobjective Decision-Making and Fuzzy Optimization Theory
The design tries to solve the problem of low pass rate of platinum wire production and the waste of platinum in company. The paper uses multiobjective decision system fuzzy optimization theory to analyze five parameters, which are tensile strength, ductility, fracture load, filling aperture, and resistance. Besides, MATLAB software is used to write programs and calculate. To sum up the above analysis, the weight vector of five parameters is obtained and that can be used to determine which parameter has the greatest influence on the pass rate of the wire winding process.
- 2 June 2019
- Published by Wiley in Journal of Control Science and Engineering
- Vol. 2019, 1-8
Open AccessDesign of Winding Parameters Based on Multiobjective Decision-Making and Fuzzy Optimization Theory
The design tries to solve the problem of low pass rate of platinum wire production and the waste of platinum in company. The paper uses multiobjective decision system fuzzy optimization theory to analyze five parameters, which are tensile strength, ductility, fracture load, filling aperture, and resistance. Besides, MATLAB software is used to write programs and calculate. To sum up the above analysis, the weight vector of five parameters is obtained and that can be used to determine which parameter has the greatest influence on the pass rate of the wire winding process.