基于掩码自编码器的干扰波形生成方法OA

In complex communication environments,existing jamming methods often result in limited increases in Bit Error Rate(BER)for target communication signals,making them insufficient for precise and effective interference in practical applications.To address this issue,this paper proposes a jamming waveform generation method based on a masked autoencoder.The proposed model employs a masking mechanism to learn the key features of the target signal,and a decoder reconstructs a structurally similar waveform.A complex-valued convolutional neural network is adopted as the backbone of the encoder to preserve the original amplitude and phase information of the signal,thereby enabling accurate feature extraction from a single target signal.Furthermore,small perturbations are injected into 1%of randomly selected non-core features,allowing the reconstructed waveform to be similar in structure but different in detail,enhancing its jamming effectiveness.This method enables targeted and efficient interference against a single signal without relying on high-power noise transmission.Simulation results show that,under the same Signal-to-Interference Ratio(SIR),the proposed method increases the BER of the target communication system by an average of 9%compared with traditional Gaussian noise jamming,demonstrating better interference performance,reduced resource waste,and improved electronic countermeasure capability.