Method for multi-functional radar operational mode recognition under small-sample conditions
In increasingly complex electromagnetic environments,the recognition of multifunction radar oper-ating modes continues to face numerous challenges.In particular,the limited number of intercepted signal samples from multifunction radars,combined with poor sample augmentation quality,leads to low accuracy in operating mode recognition.This approach is driven by the synergy of Adaptive Padding TransGAN(Gen-erative Adversarial Network with Adaptive Padding)and Model-Agnostic Meta-Learning.Initially,The Adaptive Padding TransGAN is employed for adaptive sample padding and sample augmentation,starting with the context of fitting small-sample data.Subsequently,the Model-Agnostic Meta-Learning algorithm in meta-learning is integrated to achieve precise identification of multifunctional radar operational modes under limited sample conditions.Finally,compared to algorithms combining Generative Adversarial Networks with Meta-Learning and traditional Support Vector Machine classifiers,simulation results demonstrate that the proposed approach significantly enhances recognition accuracy by 2.39%and 17.42%,respectively.The ef-fectiveness of this method in accurately identifying multifunction radar operating modes under small sample conditions has been validated.
国家科技期刊平台
NSTL
万方数据
