DRFM-Based Jamming Suppression via Convolutional Neural Network with Dilated Mechanism

In modern electronic warfare, active deception based on Digital Radio Frequency Memory (DRFM) technology poses significant challenges to radar detection. Specifically, countering mainlobe DRFM jamming is a critical difficulty in the field of radar anti-jamming. Traditional blind source separation (BSS) algorithms experience a sharp decline in jamming suppression performance when the direction of arrival of the mainlobe jamming coincides with that of the target echo. To address this issue, we have designed a deep convolutional separation network with a dilated mechanism, leveraging the principles of blind source separation. By introducing the dilated mechanism, we expand the receptive field of the convolution, enhancing the network's ability to capture the temporal dependencies of the input signals. The network parameters are updated by minimizing the mean squared error between the recovered time-domain signals and the labeled signals. The effectiveness of the proposed method is validated through experiments. The experimental results demonstrate that our proposed method not only effectively suppresses stand-off jamming but also outperforms traditional blind source separation algorithms, such as JADE, in suppressing self-defense DRFM jamming.