Jun Gui
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
Jun Gui has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Yanbin Hou
Shenyang University of Technology
3 shared publications - Linjun Si
Power China Huadong Engineering Corporation Limited, Hangzhou 310000, China
2 shared publications - Shuai Qin
School of Information Science and Engineering, Shenyang University of Technology, No. 111, Shenliao West Road, Shenyang 110870, China
2 shared publications - YuanHai Si
PowerChina Huadong Engineering Corporation Limited, No. 201, Gaojiao Road, Yuhang District, Hangzhou 311122, China
2 shared publications Baowen Li
Univ Colorado, Dept Mech Engn, Boulder, CO 80309 USA
1 shared publication- Bin Hu
Kean University
1 shared publication
Latest Publications
- 1 December 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (12) , 14132-14139
An Untethered Robotic Fish With Distributed Electromagnetic Muscle-Inspired Actuation for Multimodal Swimming
- 1 December 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (12) , 14132-14139
An Untethered Robotic Fish With Distributed Electromagnetic Muscle-Inspired Actuation for Multimodal Swimming
- Open Access
A Data-Driven Framework for Charging Voltage–Capacity Curve Prediction Using Post-Discharge Voltage Relaxation
Accurate monitoring of battery capacity degradation underpins reliable battery operation. Existing methods either estimate the current-cycle maximum capacity via full charge–discharge curves or reconstruct the current full charging curve using local charging segments. Both approaches only characterize current cycle degradation: the former suffers from unavailable full curves in practice and insufficient state description, while the latter cannot predict the aging trend of subsequent cycles. To address the above limitations, this work adopts easily collected post-discharge relaxation voltage as the core input to predict the full voltage–capacity (V-Q) curve of the next charging cycle. An LSTM-CNN hybrid feature embedding module is integrated into a Transformer-based LCformer model to combine temporal dependency modeling and local feature extraction. Furthermore, a novel customized loss function incorporating voltage-interval weighting, a monotonicity constraint, and a smoothness penalty is proposed to improve the plausibility and accuracy of the predicted curves. The proposed method achieves the prediction of a complete charging curve containing rich battery state information for the next cycle, and can provide prospective data support for relevant research based on full charging curves, such as battery state-of-health estimation and lithium plating analysis. Validation on datasets from different laboratories demonstrates that the proposed method achieves a root-mean-square error of less than 0.03 Ah in V-Q curve prediction.
Open AccessA Data-Driven Framework for Charging Voltage–Capacity Curve Prediction Using Post-Discharge Voltage Relaxation
Accurate monitoring of battery capacity degradation underpins reliable battery operation. Existing methods either estimate the current-cycle maximum capacity via full charge–discharge curves or reconstruct the current full charging curve using local charging segments. Both approaches only characterize current cycle degradation: the former suffers from unavailable full curves in practice and insufficient state description, while the latter cannot predict the aging trend of subsequent cycles. To address the above limitations, this work adopts easily collected post-discharge relaxation voltage as the core input to predict the full voltage–capacity (V-Q) curve of the next charging cycle. An LSTM-CNN hybrid feature embedding module is integrated into a Transformer-based LCformer model to combine temporal dependency modeling and local feature extraction. Furthermore, a novel customized loss function incorporating voltage-interval weighting, a monotonicity constraint, and a smoothness penalty is proposed to improve the plausibility and accuracy of the predicted curves. The proposed method achieves the prediction of a complete charging curve containing rich battery state information for the next cycle, and can provide prospective data support for relevant research based on full charging curves, such as battery state-of-health estimation and lithium plating analysis. Validation on datasets from different laboratories demonstrates that the proposed method achieves a root-mean-square error of less than 0.03 Ah in V-Q curve prediction.
- Open Access
- 4 April 2026
- Published by Elsevier in SSRN Electronic Journal
A Data-Driven Framework for Charging voltage-capacity curve Prediction Using Post-Discharge Voltage Relaxation
Accurate monitoring of battery capacity degradation is essential for ensuring stable and reliable operation. However, mainstream methods typically rely on fully available charge-discharge profiles, which are difficult to obtain in practice, to estimate the maximum capacity at the current cycle. Such approaches not only fail to comprehensively characterize the extent of battery degradation but also cannot predict the state of the battery in future cycles.To address the above challenges, this study employs the post-discharge relaxation voltage—which is easily accessible in real-world scenarios and effectively characterizes battery states—as the core input. A novel method is proposed to predict the full voltage-capacity (V-Q) curve of the subsequent charging cycle using the relaxation voltage after the current discharge cycle. We introduce a feature embedding module integrating LSTM and multi-scale CNN, leveraging their complementary strengths in modeling temporal dependencies and extracting local features to enhance the feature embedding capability of Transformer for time-series data. By combining this feature embedding module with the Transformer architecture, the LCformer model is constructed.Compared with the maximum capacity estimation method, the proposed prediction method based on LCformer can not only achieve accurate prediction of the full charging curve for the next charging cycle, but also derive richer battery state information, providing more comprehensive support for battery state evaluation. Validated on the battery dataset from Tongji University, the proposed method yields predicted V-Q curves with a root-mean-square error of less than 0.03 Ah.
- 4 April 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Data-Driven Framework for Charging voltage-capacity curve Prediction Using Post-Discharge Voltage Relaxation
Accurate monitoring of battery capacity degradation is essential for ensuring stable and reliable operation. However, mainstream methods typically rely on fully available charge-discharge profiles, which are difficult to obtain in practice, to estimate the maximum capacity at the current cycle. Such approaches not only fail to comprehensively characterize the extent of battery degradation but also cannot predict the state of the battery in future cycles.To address the above challenges, this study employs the post-discharge relaxation voltage—which is easily accessible in real-world scenarios and effectively characterizes battery states—as the core input. A novel method is proposed to predict the full voltage-capacity (V-Q) curve of the subsequent charging cycle using the relaxation voltage after the current discharge cycle. We introduce a feature embedding module integrating LSTM and multi-scale CNN, leveraging their complementary strengths in modeling temporal dependencies and extracting local features to enhance the feature embedding capability of Transformer for time-series data. By combining this feature embedding module with the Transformer architecture, the LCformer model is constructed.Compared with the maximum capacity estimation method, the proposed prediction method based on LCformer can not only achieve accurate prediction of the full charging curve for the next charging cycle, but also derive richer battery state information, providing more comprehensive support for battery state evaluation. Validated on the battery dataset from Tongji University, the proposed method yields predicted V-Q curves with a root-mean-square error of less than 0.03 Ah.
- Open Access
- 26 March 2020
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 15 (3) , e0230592
Dynamics of calling activity to toll-free numbers in China
The quantitative understanding of human behavior is a key issue in modern science. Recently, inhomogeneous human activities have been described by bursts (consecutive activities separated by long periods of inactivity) and characterized by fat-tailed inter-event time (interval between two activities) distributions. However, the dynamics between number of activities and activity duration are still unclear. In this study, we analyzed 133 million toll-free call records from China to study the dynamics between call frequency and call duration. We confirmed that both call frequency and call duration exhibit circadian cycles and weekly cycles. By analyzing intraday patterns of these two metrics, we found the opposite volatility and clustered distributions. Results of clustering analysis showed that calling activity to toll-free numbers can be clustered into four clusters. In the “Work” cluster, the distribution of call duration was significantly different from that in the other clusters. The corresponding time of “Work” cluster was much shorter than estimates based on common sense. Intraday patterns and clustering results showed that both call frequency and call duration are primarily related to circadian cycles, the nature of human beings, and that work is a secondary factor that affects these variables. Moreover, we found a strong positive correlation between call frequency and call duration, as well as polarization of joint probability. The polarization indicates two extremes in inhomogeneous calling activity to toll-free numbers, i.e., either people are very busy or very idle. The empirical probability of the extreme was approximately four times that of random probability. Our findings may have great usage for studying the dynamics of inhomogeneous human behavior.
- 26 March 2020
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 15 (3) , e0230592
Open AccessDynamics of calling activity to toll-free numbers in China
The quantitative understanding of human behavior is a key issue in modern science. Recently, inhomogeneous human activities have been described by bursts (consecutive activities separated by long periods of inactivity) and characterized by fat-tailed inter-event time (interval between two activities) distributions. However, the dynamics between number of activities and activity duration are still unclear. In this study, we analyzed 133 million toll-free call records from China to study the dynamics between call frequency and call duration. We confirmed that both call frequency and call duration exhibit circadian cycles and weekly cycles. By analyzing intraday patterns of these two metrics, we found the opposite volatility and clustered distributions. Results of clustering analysis showed that calling activity to toll-free numbers can be clustered into four clusters. In the “Work” cluster, the distribution of call duration was significantly different from that in the other clusters. The corresponding time of “Work” cluster was much shorter than estimates based on common sense. Intraday patterns and clustering results showed that both call frequency and call duration are primarily related to circadian cycles, the nature of human beings, and that work is a secondary factor that affects these variables. Moreover, we found a strong positive correlation between call frequency and call duration, as well as polarization of joint probability. The polarization indicates two extremes in inhomogeneous calling activity to toll-free numbers, i.e., either people are very busy or very idle. The empirical probability of the extreme was approximately four times that of random probability. Our findings may have great usage for studying the dynamics of inhomogeneous human behavior.
- Open Access
- 1 July 2019
- Published by Elsevier in Physica A: Statistical Mechanics and its Applications
- Vol. 525, 1343-1350
New dynamics between volume and volatility
- 1 July 2019
- Published by Elsevier in Physica A: Statistical Mechanics and its Applications
- Vol. 525, 1343-1350
Open AccessNew dynamics between volume and volatility