BRSR-OpGAN: Blind Radar Signal Restoration using Operational Generative Adversarial Network

TL;DR

The paper tackles blind restoration of radar signals corrupted by a mix of artifacts without prior artifact-type assumptions. It introduces BRSR-OpGAN, a 1D Self-Organized Operational Neural Network–based GAN framework with a dual-domain loss that combines time-domain and spectrogram-based objectives, enabling robust restoration on complex real-world signals. The authors present a compact generator–discriminator architecture, a two-pass restoration variant, and a new extended BRSR benchmark dataset, achieving state-of-the-art restoration metrics (e.g., average SNR improvements up to ~12.4 dB with 2nd-pass) and real-time CPU feasibility. This work provides a practical, scalable approach for high-fidelity radar signal recovery with potential impact on surveillance, EW, and remote sensing applications.

Abstract

Objective: Many studies on radar signal restoration in the literature focus on isolated restoration problems, such as denoising over a certain type of noise, while ignoring other types of artifacts. Additionally, these approaches usually assume a noisy environment with a limited set of fixed signal-to-noise ratio (SNR) levels. However, real-world radar signals are often corrupted by a blend of artifacts, including but not limited to unwanted echo, sensor noise, intentional jamming, and interference, each of which can vary in type, severity, and duration. This study introduces Blind Radar Signal Restoration using an Operational Generative Adversarial Network (BRSR-OpGAN), which uses a dual domain loss in the temporal and spectral domains. This approach is designed to improve the quality of radar signals, regardless of the diversity and intensity of the corruption. Methods: The BRSR-OpGAN utilizes 1D Operational GANs, which use a generative neuron model specifically optimized for blind restoration of corrupted radar signals. This approach leverages GANs' flexibility to adapt dynamically to a wide range of artifact characteristics. Results: The proposed approach has been extensively evaluated using a well-established baseline and a newly curated extended dataset called the Blind Radar Signal Restoration (BRSR) dataset. This dataset was designed to simulate real-world conditions and includes a variety of artifacts, each varying in severity. The evaluation shows an average SNR improvement over 15.1 dB and 14.3 dB for the baseline and BRSR datasets, respectively. Finally, even on resource-constrained platforms, the proposed approach can be applied in real-time.

Figures (7)

• Figure 1: Sample signals from the extended dataset demonstrating the impact of various artifacts and the effectiveness of the proposed restoration method. In each example, the clean signal (blue) and the corrupted signal (red) are overlaid in the top plot (a), while the restored signal (green) is shown in the bottom plot (b). Common artifacts include Additive White Gaussian Noise (AWGN), Echo, Interference, and a blend of all with randomized weights.
• Figure 2: The proposed blind radar restoration approach using BRSR-OpGAN. After training, the encoder and decoder networks are employed together as a composite filter $G = f_D \cdot f_E$ for denoising unseen radar signals.
• Figure 3: The generator and discriminator architectures of the BRSR-OpGAN. The ResDownBlock, ResUpBlock, and ClassificationBlock components are also detailed.
• Figure 4: The illustration of extended real-world BRSR dataset generation with a random choice of artifacts and their severities.
• Figure 5: Joint distribution of corrupted vs restored SNR.