Kaiji Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kaiji Wang has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Le Xu
National Key Laboratory of Radar Detection and Sensing, Xidian University, Xian, China
4 shared publications - Xiaowei Shi
Xidian University
2 shared publications Haitian An
Department of Electronic Engineering, National Key Laboratory of Radar Detection and Sensing, Xidian University, Xian, China
1 shared publicationHua Ji Zhou
National Key Laboratory of Electromagnetic Space Security, Jiaxing, China
1 shared publicationJincong Huo
Xian Institute of Electromechanical Information Technology, Xian, China
1 shared publication- Yao He
Department of Electronic Engineering, National Key Laboratory of Radar Detection and Sensing, Xidian University, Xian, China
1 shared publication
Latest Publications
- 11 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. PP (0018926X) , 1
A Deep Learning Framework Based on PointNet++ for RCS Prediction of Aircraft Target
In this paper, we propose RCS-PointNet++, a novel deep learning framework for predicting the radar cross section (RCS) of complex 3D structures. Built upon the PointNet++ architecture, the model effectively captures both local and global geometric features from point clouds, enabling accurate and efficient RCS prediction. RCS-PointNet++ demonstrates strong robustness to variations in point cloud resolution and generalizes well across different target configurations. Inference on 25 aircraft models across 360 directions completes in under 52 seconds. Extensive experiments on two independent datasets comprising over 2,125 3D models validate the proposed method’s accuracy and generalizability, achieving a root mean square error (RMSE) of 3.09 and an R2 of 0.92. Furthermore, enhancing the network’s perception of scattering direction reduces prediction error by approximately 6%. While initially designed for single-frequency prediction, the framework can be extended to multi-frequency scenarios by incorporating electromagnetic priors. This work offers a significant advancement in fast and accurate RCS prediction for complex 3D targets.
- 11 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. PP (0018926X) , 1
A Deep Learning Framework Based on PointNet++ for RCS Prediction of Aircraft Target
In this paper, we propose RCS-PointNet++, a novel deep learning framework for predicting the radar cross section (RCS) of complex 3D structures. Built upon the PointNet++ architecture, the model effectively captures both local and global geometric features from point clouds, enabling accurate and efficient RCS prediction. RCS-PointNet++ demonstrates strong robustness to variations in point cloud resolution and generalizes well across different target configurations. Inference on 25 aircraft models across 360 directions completes in under 52 seconds. Extensive experiments on two independent datasets comprising over 2,125 3D models validate the proposed method’s accuracy and generalizability, achieving a root mean square error (RMSE) of 3.09 and an R2 of 0.92. Furthermore, enhancing the network’s perception of scattering direction reduces prediction error by approximately 6%. While initially designed for single-frequency prediction, the framework can be extended to multi-frequency scenarios by incorporating electromagnetic priors. This work offers a significant advancement in fast and accurate RCS prediction for complex 3D targets.
- 1 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 73 (11) , 9499-9508
An Accelerated PO for Fast RCS Prediction of Electrically Large Objects
This paper proposes an accelerated physical optics (PO) method to efficiently calculate monostatic radar cross section (RCS) of electrically large objects. The shadowing test of objects is one of the most time-consuming processes of PO method. In traditional shadowing methods, the triangular facets used to describe an object are small and numerous, and each facet needs to perform occlusion judgment with all other facets. A large number of occlusion judgments render traditional methods inefficient. To improve the computational efficiency of the PO method, we propose a facet grouping method based on the scan line technique, termed the scan line grouping method (SLGM), and combine it with facet subdivision technique to accelerate the shadowing test of objects. The proposed method can rapidly group facets through SLGM, reducing the number of occlusion judgments and thereby significantly improving the computational efficiency of the PO method. The simulation results of several experiments demonstrate the effectiveness of the proposed method.
- 1 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 73 (11) , 9499-9508
An Accelerated PO for Fast RCS Prediction of Electrically Large Objects
This paper proposes an accelerated physical optics (PO) method to efficiently calculate monostatic radar cross section (RCS) of electrically large objects. The shadowing test of objects is one of the most time-consuming processes of PO method. In traditional shadowing methods, the triangular facets used to describe an object are small and numerous, and each facet needs to perform occlusion judgment with all other facets. A large number of occlusion judgments render traditional methods inefficient. To improve the computational efficiency of the PO method, we propose a facet grouping method based on the scan line technique, termed the scan line grouping method (SLGM), and combine it with facet subdivision technique to accelerate the shadowing test of objects. The proposed method can rapidly group facets through SLGM, reducing the number of occlusion judgments and thereby significantly improving the computational efficiency of the PO method. The simulation results of several experiments demonstrate the effectiveness of the proposed method.
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Antennas and Wireless Propagation Letters
- Vol. 24 (11) , 4452-4456
A Novel Adaptive Beamforming Method Based on Near-Field Scattering Coefficient Compensation
In this letter, a new adaptive beamforming method based on near-field scattering coefficient compensation (SCC) is proposed to solve the problem of rapid performance degradation of traditional beamforming methods under near-field scattering effects. The basic idea is to shift from suppressing near-field scattering signals to utilizing them. The proposed method first utilizes array received data to estimate the near-field scattering coefficient. Then, the estimated scattering coefficient is compensated to derive the optimal weight vector for beamforming. This method fully utilizes the target information in the near-field scattering multipath signals. Multiple input signals are coherently superimposed, which can greatly improve the signal-to-noise ratio (SNR) of the output signal. The results of several experiments designed indicate that the proposed method is an effective method.
- 27 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Antennas and Wireless Propagation Letters
- Vol. 24 (11) , 4452-4456
A Novel Adaptive Beamforming Method Based on Near-Field Scattering Coefficient Compensation
In this letter, a new adaptive beamforming method based on near-field scattering coefficient compensation (SCC) is proposed to solve the problem of rapid performance degradation of traditional beamforming methods under near-field scattering effects. The basic idea is to shift from suppressing near-field scattering signals to utilizing them. The proposed method first utilizes array received data to estimate the near-field scattering coefficient. Then, the estimated scattering coefficient is compensated to derive the optimal weight vector for beamforming. This method fully utilizes the target information in the near-field scattering multipath signals. Multiple input signals are coherently superimposed, which can greatly improve the signal-to-noise ratio (SNR) of the output signal. The results of several experiments designed indicate that the proposed method is an effective method.