Machine Learning-Based Frequency Extrapolation of Radar Cross Section for a Complex Target

Radar cross section (RCS), as a critical parameter of radar targets, plays an indispensable role in both engineering and military applications. This paper investigates the prediction of RCS for complex targets using five machine learning-based methods, where the original training samples are generated through full-wave electromagnetic simulations. The proposed prediction approaches achieve promising performance and alleviate the high computational cost and time consumption associated with traditional RCS calculations. A series of experiments on a complex tank target are conducted to evaluate the accuracy and effectiveness of the RCS prediction models. The results demonstrate that the five machine learning methods offer high precision in frequency-domain RCS extrapolation and exhibit strong robustness, providing a new perspective for future research on RCS sequence prediction of targets.