Abstract:

Aiming at the problems of difficult target signal identification and weakened jamming effectiveness of traditional electronic jamming methods in complex electromagnetic environment, a jamming signal generation model based on self-attentive boundary-balanced generative adversarial network is proposed. The proposed model uses the up-sampling module to enhance the fineness of the generated data, and introduces the self-attention mechanism to take into account the local and global nature of the signal feature extraction, which not only makes the results more accurate, but also reduces the computational complexity. A discriminator based on the self-encoder architecture is also used to facilitate the fast and stable convergence of the model. The experimental results show that the proposed model can adaptively identify and learn the non-cooperative target signals and automatically generate the corresponding jamming signals, and the jamming efficiency is better than the traditional jamming algorithms and the classical generative adversarial network model algorithms, which provides a new research idea for the communication adversarial technology based on machine learning.

Key words: communication countermeasures, generative adversarial network, self-attention mechanism, self-encoder, signal generation