5GNRad: 5G New Radio Radar

MATLAB end to end radar processing chain for 5G NR.

5GNRad is a MATLAB-based simulation platform to evaluate sensing with cellular infrastructure using standard-compliant 5G NR Positioning Reference Signals (PRS) and a radar-style processing chain that produces range–Doppler–angle detections and target position/velocity estimates.

---

Key features

• Standard-compliant PRS waveform generation (NR numerology + OFDM mapping)
• Two operating modes
  1. Uploaded MPC channels (3GPP Release 19, TR 38.901 sensing extensions) via backgroundChannel.json / targetChannel.json
  2. Simplified single-point target (single scatterer) with RCS-based amplitude scaling aligned with TR 38.901 sensing formulations
• Radar processing chain
  • Range processing (IFFT / delay profile)
  • Doppler processing (slow-time FFT)
  • Angle estimation (beamspace FFT or Bartlett scan, depending on config)
  • Detection: 2D CFAR + peak picking + clustering
• Scenario-driven: each example is self-contained under examples/<scenarioName>/Input/
• Parallel execution over drops (parfor) when enabled

Requirements

Validated on MATLAB R2025a.

Required MathWorks products

• MATLAB
• 5G Toolbox (NR numerology, PRS generation, OFDM grid mapping)
• Phased Array System Toolbox (arrays, steering vectors, angle utilities)
• Signal Processing Toolbox (windowing and FFT helpers)

Optional

• Parallel Computing Toolbox (parallel execution across drops)

---

Repository structure

<repo-root>/
  main.m
  setup.m
  src/
    +nrRadar/
      +cfg/        (scenario configuration readers/validators)
      +rx/         (RX synthesis, OFDM demod, RD/RDA formation)
      +sens/       (CFAR, NMS, DOA, clustering, sidelobe suppression, geometry)
      +array/      (array utilities and coordinate conventions)
      +io/         (export utilities)
      +internal/   (precompute + per-drop worker entry points)
  examples/
    <scenarioName>/
      Input/       (configs + optional channel JSONs)
      Output/      (generated results)

---

Quick start

From the repository root:

setup();
main

main.m runs a batch of scenarios listed inside the script (config → precompute → per-drop processing → export).

Run a single scenario

setup();

scenarioPath = "examples/uma_trp1_3gpp";

[simConfig, stConfig, prsConfig, geometry, sensConfig, backgroundChannel, targetChannel] = ...
    nrRadar.cfg.configureScenario(scenarioPath);

desiredWorkers = 8;          % optional
parallelMode   = "off";      % "on" | "off" 

[results, detStats, detectionOutput] = nrRadar.run( ...
    simConfig, stConfig, prsConfig, geometry, sensConfig, ...
    backgroundChannel, targetChannel, desiredWorkers, ...
    "Parallel", parallelMode);

nrRadar.io.exportResults(scenarioPath, results, detStats, detectionOutput);

---

Scenario inputs

Each scenario under examples/<scenarioName>/Input/ includes:

• simulationConfig.txt — carrier, bandwidth, noise figure, scenario options, etc.
• prsConfig.txt — PRS resource definition (periodicity, comb size, symbol mapping, etc.)
• sensingConfig.txt — processing and detection settings (FFT sizes, CFAR, DOA method, thresholds, clustering)
• txAntennaModel.json, rxAntennaModel.json — array geometry + beamforming/combining configuration
• backgroundChannel.json (optional) — environment/clutter MPCs (uploaded-channel mode)
• targetChannel.json (optional) — target MPCs (uploaded-channel mode)
• bsConfiguration.txt — TRP/BS configuration
• targetConfiguration.txt — target truth states (position/velocity timeline)

Operating mode selection

• If targetChannel.json / backgroundChannel.json exist, the simulator uses uploaded MPC mode.
• If target MPCs are omitted (or the scenario is configured accordingly), the simulator can run in simplified single-point target mode (single scatterer with RCS-based amplitude scaling).

Outputs

After a run, results are written under:

examples/<scenarioName>/Output/

Outputs include:

• detection.json — per-drop detection payloads (estimated positions/velocities and related metrics)
• detStats.csv — per-drop detection statistics (TP / FN / FP / false-alarm probability)
• error.csv — per-target/per-drop error table (position/range/angle/velocity errors; NaN if unassigned)

---

Performance tips

• Use "Parallel","on" with a sensible desiredWorkers to speed up batch runs.
• Runtime is dominated by:
  • number of drops / snapshots
  • range/Doppler/angle FFT sizes
  • number of RF chains (digital channels)
  • MPC count and channel length
• If you hit out-of-memory errors, reduce worker count.

---

Documentation

Documentation is available in the docs/5G_NR_Radar_doc.pdf file.

---

Publications and 3GPP Contributions

• S. Blandino et al., “Detecting Airborne Objects with 5G NR Radars,” IEEE MILCOM 2025.

• National Institute of Standards and Technology (NIST), “R1-2505684 Discussion on ISAC Performance Evaluation,” 3GPP TSG-RAN WG1 Meeting #122

• National Institute of Standards and Technology (NIST), “R1-2508883 Discussion on ISAC Performance Evaluation,” 3GPP TSG-RAN WG1 Meeting #123

• National Institute of Standards and Technology (NIST), “R1-2600846 Discussion on ISAC Performance Evaluation,” 3GPP TSG-RAN WG1 Meeting #124

---

Contributing

Feedback and contributions are welcome! Please open an issue or contact the maintainer.

Contact

Steve Blandino
NIST Communications Technology Laboratory
steve.blandino@nist.gov
https://www.nist.gov/people/steve-blandino

Citing 5GNRad

If you use 5GNRad to generate figures or results for a publication or standards contribution, please cite the associated manuscript and the software repository.

Manuscript

• S. Blandino et al., "Detecting Airborne Objects with 5G NR Radars," MILCOM 2025 - 2025 IEEE Military Communications Conference (MILCOM), Los Angeles, CA, USA, 2025, pp. 1260-1265, doi: 10.1109/MILCOM64451.2025.11310754.

Software

• National Institute of Standards and Technology (NIST) Communications Technology Laboratory, 5G New Radio Radar (5GNRad) [Computer software]. Available: https://github.com/usnistgov/5GNRad

BibTeX

@INPROCEEDINGS{11310754,
  author={Blandino, Steve and Golmie, Nada and Sahoo, Anirudha and Nguyen, Thao and Ropitault, Tanguy and Griffith, David and Sonny, Amala},
  booktitle={MILCOM 2025 - 2025 IEEE Military Communications Conference (MILCOM)}, 
  title={Detecting Airborne Objects with 5G NR Radars}, 
  year={2025},
  volume={},
  number={},
  pages={1260-1265},
  keywords={Three-dimensional displays;Radar clutter;Simulation;Airborne radar;Urban areas;Radar detection;Estimation;Autonomous aerial vehicles;Integrated sensing and communication;Clutter;ISAC;5G mobile communication;6G mobile communication;3GPP Standards;Target Detection},
  doi={10.1109/MILCOM64451.2025.11310754}}

@software{nist_5gnrad,
  author  = {{National Institute of Standards and Technology (NIST) Communications Technology Laboratory}},
  title   = {5G New Radio Radar (5GNRad)},
  url     = {https://github.com/usnistgov/5GNRad},
  note    = {Computer software}
}