TY - GEN
T1 - Radar Target Recognition Based on Electromagnetic Scattering Characteristics and Simulated Echo Data
AU - Liu, Chang
AU - Cui, Lele
AU - Wang, Zhifa
AU - Zhang, Zien
AU - Jin, Zeyu
AU - Zhang, Guangwei
N1 - Publisher Copyright:
© 2026 Global IT Research Institute - GIRI.
PY - 2026
Y1 - 2026
N2 - In complex electromagnetic environments, noise, rough-surface scattering, and clutter severely degrade radar target recognition performance. To enhance recognition capability under low Signal-to-Noise Ratio (SNR) conditions, this paper proposes a radar target recognition method that integrates electromagnetic scattering modelling, CST-based Radar Cross Section (RCS) simulation, and an improved one-dimensional Convolutional Neural Network (1D CNN). High-fidelity Linear Frequency-Modulated (LFM) echo signals are constructed from CST-based multi-angle RCS data. A wavelet transform module, a convolutional denoising block, and the Transformer are introduced to strengthen multi-scale feature representation and noise robustness. Experimental results demonstrate that the proposed model outperforms conventional CNN methods in terms of accuracy and F1-score, and maintains a recognition rate of 93% in low-SNR scenarios, indicating strong robustness and practical applicability.
AB - In complex electromagnetic environments, noise, rough-surface scattering, and clutter severely degrade radar target recognition performance. To enhance recognition capability under low Signal-to-Noise Ratio (SNR) conditions, this paper proposes a radar target recognition method that integrates electromagnetic scattering modelling, CST-based Radar Cross Section (RCS) simulation, and an improved one-dimensional Convolutional Neural Network (1D CNN). High-fidelity Linear Frequency-Modulated (LFM) echo signals are constructed from CST-based multi-angle RCS data. A wavelet transform module, a convolutional denoising block, and the Transformer are introduced to strengthen multi-scale feature representation and noise robustness. Experimental results demonstrate that the proposed model outperforms conventional CNN methods in terms of accuracy and F1-score, and maintains a recognition rate of 93% in low-SNR scenarios, indicating strong robustness and practical applicability.
KW - CNN
KW - LFM Radar Echo
KW - RCS Simulation
KW - Radar Target Recognition
KW - Target Scattering Characteristics
UR - https://www.scopus.com/pages/publications/105036092169
U2 - 10.23919/ICACT68090.2026.11431566
DO - 10.23919/ICACT68090.2026.11431566
M3 - Conference contribution
AN - SCOPUS:105036092169
T3 - International Conference on Advanced Communication Technology, ICACT
SP - 246
EP - 251
BT - 28th International Conference on Advanced Communications Technology
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 28th International Conference on Advanced Communications Technology, ICACT 2026
Y2 - 8 February 2026 through 11 February 2026
ER -