MREDNet: A Multi-Residual Encoder-Decoder Network for Recovering LFM Interfered by Intra-Pulse Retransmission Interferences
openalex(2023)
Abstract
Retransmission interference in radar and communication signal processing significantly threatens the received linear frequency modulation signal. With the advancement in the hardware capabilities of interference devices, they can intercept and retransmit signals within the duration of a single pulse. Therefore, we refer to this type of interference as intra-pulse retransmission interference, which includes retransmission modes smeared spectrum, interrupted sampling repeater jamming, and interrupted nonuniform sampling repeater jamming. Although traditional interference cancellation techniques offer optimal solutions under the assumption of fully known interference parameters, practical scenarios often involve unknown parameters, estimation errors, random parameters, and multiple interferences, leading to performance degradation. To address these problems, we propose a Multi-Residual Encoder-Decoder Network (MRDENet), a novel approach for effectively mitigating intra-pulse retransmission interferences and recovering the signal of linear frequency modulation. In contrast to parameter estimation and interference cancellation, MREDNet extracts crucial features of both the target and interference and performs reconstruction. More specifically, MREDNet acquires information about the retransmission pattern of interference and the modulation frequency through its multi-level residual shortcut connections. To address the issue of network parameters being real while the signals are complex, we devised decomplexification—a data preprocessing approach enabling neural networks for direct utilization in front-end complex signal processing. Besides, we evaluate the performance of MREDNet comprehensively. The experimental conditions encompassed single and multiple interferences, incomplete data sets, varying signal-to-noise ratios, and jamming-to-signal ratios. The computational simulations demonstrate that MREDNet successfully solves the problem of recovering linear frequency modulation signals interfered with by intra-pulse retransmission interference. According to the boundary condition of MREDNet, the proposed method exhibits valuable applications in real-world scenarios.
MoreTranslated text
Key words
Signal processing,Intra-pulse retransmission interference,Smeared spectrum,Interrupted sampling repeater jamming,Interrupted nonuniform sampling repeater jamming,Unknown interference parameters,Signal reconstruction,Deep learning,Interference cancellation
AI Read Science
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined