基于生成对抗网络的伪码调相脉冲多普勒引信引导式干扰方法

针对低信噪比环境下难以精准估计复合调制引信信号参数，造成传统引导式干扰方法对伪码调相脉冲多普勒引信干扰效率低的问题，提出基于生成对抗网络（GAN）的伪码调相脉冲多普勒引信引导式干扰方法，实现低信噪比环境下对伪码调相脉冲多普勒引信的有效干扰。对伪码调相脉冲多普勒引信下变频获取中频信号并预处理，获取去噪处理后的时频分布及相位等信息。通过GAN模型利用时频分布数据进行数据生成，在低信噪比环境下无需估计引信参数即可生成与引信中频信号相关度极高的数据。将获得的数据利用逆变换生成时域波形重构出引信干扰信号。实验结果表明，在低信噪比环境下，基于GAN的伪码调相脉冲多普勒引信引导式干扰方法能够对伪码调相脉冲多普勒引信产生良好的干扰效果。

Guided Jamming Method for Pseudo-code Phase-modulated Pulse Doppler Fuze Based on Generative Adversarial Nets

For the pseudo-code phase-modulated pulse Doppler fuzes, conventional guided jamming methods have low interference efficiency due to the difficulty of accurately estimating the compound modulation signal parameters in low SNR environments. A guided jamming method for the pseudo-code phase-modulated pulse Doppler fuze based on the generative adversarial nets (GAN) is presented to solve this problem, which can effectively interfere with the pseudo-code phase-modulated pulse Doppler fuze at low SNR. First, the intermediate frequency signal obtained by down-conversion of the pseudo-code phase-modulated pulse Doppler fuze is preprocessed. Then, the denoised information such as time-frequency distribution and phase is obtained. Second, the GAN model is used with the time-frequency distribution data to generate highly correlated data with the fuze intermediate frequency signal without having to estimate the fuze parameters in a low SNR environment. Finally, the fuze interference signal is reconstructed using the time-domain waveform generated by the inverse transformation of the newly generated data. Simulation and experiment results show that the guided jamming method based on GAN can achieve a better jamming effect on the pseudo-code phase-modulated pulse Doppler fuze in low SNR environments.