SEMUS - An Open-Source RF-Level SAR Emulator for Interference Modeling in Spaceborne Applications

Nermine Hendy, Ferdi Kurnia, Thomas Kraus, Markus Bachmann, Marco Martorella, Robin J. Evans, Manfred Zink, Haytham M. Fayek, Akram Al-HouraniIEEE Trans. Aerosp. Electron. Syst 2025 · DOI

Introduction

Synthetic Aperture Radar (SAR) has become an indispensable tool for Earth observation, providing high-resolution imagery regardless of weather conditions or daylight. By transmitting microwave pulses and processing the returned echoes, SAR systems can "see" through clouds and darkness. However, as the radio frequency (RF) spectrum becomes increasingly crowded with signals from IoT devices, terrestrial radar, and other satellite constellations, the integrity of SAR data is under threat from Radio Frequency Interference (RFI).

The Challenge of Radio Frequency Interference (RFI)

RFI manifests as artifacts or "ghosts" in the final processed SAR images, which can obscure critical features and render expensive satellite data unusable. While the community has developed various mitigation strategies, there is a distinct lack of open-source tools capable of simulating the entire physical process of SAR acquisition at the RF level. Most existing simulators are either proprietary or focus solely on processing Level-1 (Single Look Complex) images, skipping the raw "Level-0" phase where interference first interacts with the radar signal.

Real-world RFI artifacts in a SAR image, illustrating the degradation of satellite observations.

The SEMUS Framework

To bridge this gap, the authors introduce SEMUS (SAR EMUlator for Spaceborne applications), a new open-source framework developed in MATLAB. SEMUS is designed to simulate the end-to-end SAR pipeline, from orbital mechanics to image reconstruction, with a specific focus on modeling RFI at the hardware level. The framework is built upon a modular four-block architecture that allows researchers to customize every stage of the simulation.

The four-block architectural framework of the SEMUS emulator, showing the interaction between geometric, target, signal, and interference models.

Methodology

1. Geometric Model

The simulation begins by defining the satellite's orbit using Keplerian elements (such as inclination, eccentricity, and altitude). SEMUS employs a "start-and-stop" model to calculate the precise distance between the moving satellite and ground targets for every pulse, ensuring accurate phase history generation.

2. Target Model

Rather than relying on simplistic point targets, SEMUS utilizes real-world optical satellite imagery to create complex simulation environments. By interpolating the pixel intensities of these images, the emulator estimates ground reflectivity ((\sigma)), providing a realistic proxy for the radar cross-section of a scene.

3. Signal Transceiver Model

This block generates the Linear Frequency Modulated (LFM) chirps characteristic of SAR systems. It calculates the time delay and amplitude of the echoes returning from the ground reflectivity map, producing the raw "Level-0" data (phase history) that a real satellite would record.

4. Noise and Interference Model

The core contribution of SEMUS is the ability to inject arbitrary interference directly into the raw data. Users can add standard thermal noise (Additive White Gaussian Noise) or complex RFI waveforms. The paper specifically highlights the threat of Long Range (LoRa) signals, which operate in the L-band and are increasingly common in smart city IoT deployments.

Image Reconstruction and Results

For image processing, SEMUS implements the Range-Doppler Algorithm (RDA). A notable feature is the inclusion of an "empirical azimuth filter," which the authors found to provide superior focusing in certain simulated scenarios compared to traditional analytical filters.

Validation of the emulator was conducted through three primary test cases using a simulated scene of Melbourne, Australia:

  • Clean Data: The system successfully reconstructed high-fidelity images from pure simulated echoes.
  • Noise Contamination: The emulator demonstrated robustness when thermal noise was added at a Signal-to-Noise Ratio (SNR) of 24 dB.
  • RFI Injection: By injecting a LoRa signal at a Signal-to-Interference Ratio (SIR) of 32 dB, the authors were able to reproduce the distinct vertical line artifacts seen in real-world RFI cases. This qualitative match proves the emulator's utility in modeling how specific wireless technologies interfere with spaceborne sensors.

Conclusion and Future Work

SEMUS provides the research community with a much-needed open-source tool for studying RFI mitigation at the RF level. By making the code available on GitHub, the authors hope to accelerate the development of new algorithms for protecting satellite observations. Future iterations of SEMUS are expected to support more advanced imaging modes like Spotlight and ScanSAR, as well as faster processing for larger-scale geographical scenes.

Summary updated

BibTeX
@article{hendy2025semus,
  author = {Nermine Hendy and Ferdi Kurnia and Thomas Kraus and Markus Bachmann and Marco Martorella and Robin J. Evans and Manfred Zink and Haytham M. Fayek and Akram Al-Hourani},
  title = {{SEMUS - An Open-Source RF-Level SAR Emulator for Interference Modeling in Spaceborne Applications}},
  journal = {IEEE Trans. Aerosp. Electron. Syst.},
  year = {2025},
  doi = {10.1109/taes.2024.3462895},
}
Figure