Wenhao Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Wenhao Chen has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Chengjun Guo
School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China
3 shared publications - Qing Zhao
School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu, China
3 shared publications - Sicun Han
Rensselaer Polytechnic Institute, New York, US
2 shared publications - Yunhao Liu
School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu, China
2 shared publications - Li Wang
Southwest China Institute of Electronic Technology
1 shared publication Qingquan Xia
School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu, China
1 shared publication
Latest Publications
- Open Access
- 20 April 2026
- Published by Springer Nature in Journal on Advances in Signal Processing
- Vol. 2026 (1)
A new jamming detection approach with reduced false alarm probability for high-speed frequency-hopping system
- 20 April 2026
- Published by Springer Nature in Journal on Advances in Signal Processing
- Vol. 2026 (1)
Open AccessA new jamming detection approach with reduced false alarm probability for high-speed frequency-hopping system
- 2 December 2025
- Published by Springer Nature in Signal, Image and Video Processing
- Vol. 19 (17) , 1402
Intelligent radar modulation recognition with multi-node cooperation and spatial–channel attention
- 2 December 2025
- Published by Springer Nature in Signal, Image and Video Processing
- Vol. 19 (17) , 1402
Intelligent radar modulation recognition with multi-node cooperation and spatial–channel attention
- Open Access
- 27 March 2025
- Published by MDPI AG in Preprints.org
An Open-Set Recognition Method for Active Radar Jamming Signals Based on Multi-Dimensional Prototype Feature Constraints and Adversarial Reciprocal Point Learning
In the increasingly complex electromagnetic environment, accurately recognizing unknown types of active radar jamming signals has become a task of significant research importance and considerable challenge. Most current intelligent jamming recognition methods struggle to effectively handle unknown jamming signals. Although some open-set jamming recognition algorithms have achieved a certain level of discrimination for unknown jamming in recent years, the introduction of unknown jamming often severely impacts their closed-set recognition performance, especially under low Jamming-to-Noise Ratio(JNR) conditions, where the overall model performance significantly declines. To address these issues, this paper proposes an Open-Set Recognition(OSR) optimization method based on multi-dimensional prototype feature constraints and adversarial reciprocal point learning, termed MDPAR, aiming to simultaneously improve the rejection capability for unknown jamming signals and the closed-set classification accuracy for known jamming signals. First, based on prototype points, a Multi-Dimensional Prototype Feature Constraint(MDPFC) is proposed. By designing Intra-class Compactness Loss(ICL), Inter-class Boundary Loss(IBL), and open-point cosine loss, we optimized the distribution of the feature space for known classes. Second, we introduced an Adversarial margin-based Reciprocal Point Learning method(ARPL) to model the open space where unknown jamming signals reside. Additionally, an Associated Soft Constraint(ASC) between open points and reciprocal points is proposed to jointly optimize and unify the overall distribution of the open space. Finally, the jamming signal data, after domain transformation, is input into the feature extraction network, and the MDPAR model is used for training and testing. Experimental results demonstrate that the proposed MDPAR method exhibits excellent performance across multiple OSR evaluation metrics. Particularly, under conditions where the Jamming-to-Signal Ratio(JSR) is 5 dB and the Signal-to-Noise Ratio(SNR) is above 0 dB, the recognition rate for unknown jamming signals remains above 90\%. Furthermore, through threshold visualization and t-SNE dimensionality reduction analysis of the MDPAR prediction results, clear boundaries between unknown and known class samples are formed, and significant inter-class separation structures are observed within known classes, further validating the effectiveness and advancement of MDPAR in the task of OSR for radar jamming signals.
- 27 March 2025
- Published by MDPI AG in Preprints.org
Open AccessAn Open-Set Recognition Method for Active Radar Jamming Signals Based on Multi-Dimensional Prototype Feature Constraints and Adversarial Reciprocal Point Learning
In the increasingly complex electromagnetic environment, accurately recognizing unknown types of active radar jamming signals has become a task of significant research importance and considerable challenge. Most current intelligent jamming recognition methods struggle to effectively handle unknown jamming signals. Although some open-set jamming recognition algorithms have achieved a certain level of discrimination for unknown jamming in recent years, the introduction of unknown jamming often severely impacts their closed-set recognition performance, especially under low Jamming-to-Noise Ratio(JNR) conditions, where the overall model performance significantly declines. To address these issues, this paper proposes an Open-Set Recognition(OSR) optimization method based on multi-dimensional prototype feature constraints and adversarial reciprocal point learning, termed MDPAR, aiming to simultaneously improve the rejection capability for unknown jamming signals and the closed-set classification accuracy for known jamming signals. First, based on prototype points, a Multi-Dimensional Prototype Feature Constraint(MDPFC) is proposed. By designing Intra-class Compactness Loss(ICL), Inter-class Boundary Loss(IBL), and open-point cosine loss, we optimized the distribution of the feature space for known classes. Second, we introduced an Adversarial margin-based Reciprocal Point Learning method(ARPL) to model the open space where unknown jamming signals reside. Additionally, an Associated Soft Constraint(ASC) between open points and reciprocal points is proposed to jointly optimize and unify the overall distribution of the open space. Finally, the jamming signal data, after domain transformation, is input into the feature extraction network, and the MDPAR model is used for training and testing. Experimental results demonstrate that the proposed MDPAR method exhibits excellent performance across multiple OSR evaluation metrics. Particularly, under conditions where the Jamming-to-Signal Ratio(JSR) is 5 dB and the Signal-to-Noise Ratio(SNR) is above 0 dB, the recognition rate for unknown jamming signals remains above 90\%. Furthermore, through threshold visualization and t-SNE dimensionality reduction analysis of the MDPAR prediction results, clear boundaries between unknown and known class samples are formed, and significant inter-class separation structures are observed within known classes, further validating the effectiveness and advancement of MDPAR in the task of OSR for radar jamming signals.
- 2 February 2020
- Published by Springer Nature
- p. 498-508
A Rectified Linear Unit Model for Diagnosing VCSEL’s Power Output
- 2 February 2020
- Published by Springer Nature
- p. 498-508
A Rectified Linear Unit Model for Diagnosing VCSEL’s Power Output