Yu Kuang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yu Kuang has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Zhi Ning Chen
National University of Singapore
3 shared publications - Qingsha S. Cheng
Department of Electronic and Electrical Engineering, Southern University of Science and Technology, Shenzhen, China
2 shared publications - Bo Zhang
Department of Electrical and Computer Engineering, National University of Singapore, Queenstown, Singapore
1 shared publication - Danyu Yang
Department of Electrical and Computer Engineering, National University of Singapore, Queenstown, Singapore
1 shared publication - Jian-Ming Jin
Department of Electrical and Computer Engineering, University of Illinois at UrbanaChampaign, Champaign, IL, USA
1 shared publication - Peiqin Liu
Department of Electrical and Computer Engineering, National University of Singapore, Singapore
1 shared publication
Latest Publications
- Top Cited Papers
- 23 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 74 (5) , 4051-4066
Deep Learning-Enabled Generative Antenna Design: Review and Prospects
This paper explores generative antenna design (GAD) and optimization using neural network (NN)-based deep learning (DL) methods. We first outline traditional antenna design and the challenges of applying NN-DL synthesis, particularly with generative adversarial networks (GANs). While conventional approaches reach performance limits, applying NN-DL generative methods to designs with increased degrees of freedom offers new possibilities for enhanced performance, additional functionality, and knowledge discovery. We then review the historical development and diverse applications of NN-based antenna optimization, covering individual elements, multi-antennas, and periodic structures. Key algorithmic concepts such as conditional GANs and variational autoencoders (VAEs) are discussed, emphasizing the role of prior knowledge in GAD. Recent advancements demonstrate how DL-enabled synthesis can surpass phase shift limits, control dispersion, and enhance metalens antenna gain bandwidth by 52%. The approach also enables full-range amplitude and phase control, effectively reducing sidelobe levels to below −30 dB. Finally, we highlight the significance of prior knowledge in GAD, distinguishing GAD from conventional parametric optimization. We discuss the challenges, future trends, and broader impacts of NN-driven generative design in next-generation antenna and electromagnetic device engineering.
- 23 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 74 (5) , 4051-4066
Top Cited PapersDeep Learning-Enabled Generative Antenna Design: Review and Prospects
This paper explores generative antenna design (GAD) and optimization using neural network (NN)-based deep learning (DL) methods. We first outline traditional antenna design and the challenges of applying NN-DL synthesis, particularly with generative adversarial networks (GANs). While conventional approaches reach performance limits, applying NN-DL generative methods to designs with increased degrees of freedom offers new possibilities for enhanced performance, additional functionality, and knowledge discovery. We then review the historical development and diverse applications of NN-based antenna optimization, covering individual elements, multi-antennas, and periodic structures. Key algorithmic concepts such as conditional GANs and variational autoencoders (VAEs) are discussed, emphasizing the role of prior knowledge in GAD. Recent advancements demonstrate how DL-enabled synthesis can surpass phase shift limits, control dispersion, and enhance metalens antenna gain bandwidth by 52%. The approach also enables full-range amplitude and phase control, effectively reducing sidelobe levels to below −30 dB. Finally, we highlight the significance of prior knowledge in GAD, distinguishing GAD from conventional parametric optimization. We discuss the challenges, future trends, and broader impacts of NN-driven generative design in next-generation antenna and electromagnetic device engineering.
- 16 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 74 (4) , 3385-3396
Chu-Limit-Guided Decomposition-Based Multiobjective Large-Scale Optimization for Generative Broadband Electrically Small Antenna Design
In traditional electrically small antennas (ESAs) design frameworks, overcoming the bandwidth limit derived from the Chu limit is considered unachievable. This paper proposes a three-dimensional (3-D) ESA configuration with high degrees of freedom (DoF), which is employed to significantly expand the design space for beyond-limit solutions. A revised multiobjective evolutionary algorithm based on decomposition (MOEA/D) is presented, which is guided by a Pareto front derived from the bandwidth limit to achieve rapid convergence. Prior knowledge, including structural priors, fabrication constraints, and physical interpretations of design parameters and objectives, is incorporated to efficiently generate effective solutions. The proposed method generates nine solutions, all of which exceed the bandwidth limits. A “radial-beamlet” antenna, fabricated using metal 3-D printing, achieves impedance bandwidth of 50.3% (2.05-3.43 GHz) and realized gain of 1.72-3.86 dBi. The bandwidth-efficiency product of the fabricated solution is 0.402, compared to the performance limit of 0.389. This approach pushes the boundaries of electrically small antennas and provides a new pathway for the advanced design of similar structures.
- 16 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 74 (4) , 3385-3396
Chu-Limit-Guided Decomposition-Based Multiobjective Large-Scale Optimization for Generative Broadband Electrically Small Antenna Design
In traditional electrically small antennas (ESAs) design frameworks, overcoming the bandwidth limit derived from the Chu limit is considered unachievable. This paper proposes a three-dimensional (3-D) ESA configuration with high degrees of freedom (DoF), which is employed to significantly expand the design space for beyond-limit solutions. A revised multiobjective evolutionary algorithm based on decomposition (MOEA/D) is presented, which is guided by a Pareto front derived from the bandwidth limit to achieve rapid convergence. Prior knowledge, including structural priors, fabrication constraints, and physical interpretations of design parameters and objectives, is incorporated to efficiently generate effective solutions. The proposed method generates nine solutions, all of which exceed the bandwidth limits. A “radial-beamlet” antenna, fabricated using metal 3-D printing, achieves impedance bandwidth of 50.3% (2.05-3.43 GHz) and realized gain of 1.72-3.86 dBi. The bandwidth-efficiency product of the fabricated solution is 0.402, compared to the performance limit of 0.389. This approach pushes the boundaries of electrically small antennas and provides a new pathway for the advanced design of similar structures.
- 6 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 71 (12) , 9717-9727
Characteristic Mode-Guided Trust-Region-Based Optimization for Mode Manipulation in Dual-Band Metantenna Design
A characteristic mode (CM)-guided trust-region-based optimization technique for dual-band metantenna design is proposed. The technique automatically manipulates the CMs of a metantenna to achieve desired radiation patterns at desired frequency bands. Mode selection and tracking as the key steps of the technique are developed to track and detect desired, unwanted, and new unwanted modes. A trust-region optimization framework with sensitivities is exploited to manipulate the desired modes and adaptively remove unwanted modes. By optimizing the CMs of nonspecific-feeding-structure, the design achieves enhanced performance with a simplified structure. A dual-band metantenna for 5G applications, as an example, is optimized to obtain two CMs resonating at 28.4 and 38 GHz, respectively without any unwanted modes below 44 GHz. Then a feeding network is designed to excite the desired modes of metantenna. The design verifies the proposed method with the realized gain of 3.7-7.23 dBi and 5.7-10.1 dBi as well as desired consistent radiation patterns over the measured impedance bandwidths (for |S11| < -10 dB) of 22.2-28.2 GHz and 34.3-41.6 GHz, respectively.
- 6 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 71 (12) , 9717-9727
Characteristic Mode-Guided Trust-Region-Based Optimization for Mode Manipulation in Dual-Band Metantenna Design
A characteristic mode (CM)-guided trust-region-based optimization technique for dual-band metantenna design is proposed. The technique automatically manipulates the CMs of a metantenna to achieve desired radiation patterns at desired frequency bands. Mode selection and tracking as the key steps of the technique are developed to track and detect desired, unwanted, and new unwanted modes. A trust-region optimization framework with sensitivities is exploited to manipulate the desired modes and adaptively remove unwanted modes. By optimizing the CMs of nonspecific-feeding-structure, the design achieves enhanced performance with a simplified structure. A dual-band metantenna for 5G applications, as an example, is optimized to obtain two CMs resonating at 28.4 and 38 GHz, respectively without any unwanted modes below 44 GHz. Then a feeding network is designed to excite the desired modes of metantenna. The design verifies the proposed method with the realized gain of 3.7-7.23 dBi and 5.7-10.1 dBi as well as desired consistent radiation patterns over the measured impedance bandwidths (for |S11| < -10 dB) of 22.2-28.2 GHz and 34.3-41.6 GHz, respectively.