Adaptive Jamming Waveform Generation Utilizing Denoising Diffusion Probability Models

Jamming attack is a critical technique in communication countermeasures. This paper proposes a novel jamming method that employs Denoising Diffusion Proba-bilistic Models (DDPM) to generate distorted signals, which effectively increase the bit error rate (BER). The underlying mechanism functions similarly to a parroting technique, where the attacker replicates the transmission behavior during the communication process and transmits nonsensical information to cause interference. Compared to additive white Gaussian noise (AWGN) jamming, the proposed method demonstrates a more favorable energy efficiency ratio.