GlobalAmBC-DRL: SBRC 2026 Artifact

Overview

This repository provides the experimental artifact for the paper:

"A New Centralized DRL-Based Control Module for Dense Batteryless IoT Networks with Ambient Backscatter"

The goal of this artifact is to enable reproducibility of the main results presented in the paper, particularly the temporal behavior of the network success rate under different control strategies.

---

Artifact Goals

This artifact reproduces the behavior presented in Figure 3, demonstrating:

• Instability in networks without control (baseline)
• Stabilization using the GlobalAmBC-DRL module

---

Repository Structure

GlobalAmBC-DRL-SBRC2026/
│
├── data/        # Input datasets
├── scripts/     # Execution pipeline
├── drl_agent/   # DRL components (DDPG-like structure)
├── results/     # Generated outputs
├── figures/     # Final figures
├── docs/        # Documentation

---

Artifact Evaluation Badges

This artifact targets the following badges:

• Artifacts Available (SeloD)
• Artifacts Functional (SeloF)
• Results Reproduced (SeloR)

---

Basic Information

The artifact can be executed on standard environments:

• Operating System: Windows, Linux, or macOS
• Python: version 3.8 or higher
• RAM: at least 2 GB
• Disk: minimal (<100 MB)

---

Dependencies

Install the required dependencies:

pip install -r requirements.txt

Dependencies include:

• pandas
• numpy
• matplotlib

---

Installation

Clone the repository and install dependencies:

git clone https://github.com/LABORA-INF-UFG/GlobalAmBC-DRL-SBRC2026
cd GlobalAmBC-DRL-SBRC2026
pip install -r requirements.txt

---

Minimal Test

To verify that the artifact is working correctly, run:

python scripts/plot_figure3.py

Expected result:

• File figures/figure3.png is generated
• A plot comparing baseline vs GlobalAmBC-DRL is displayed

---

Experiments

Claim 1 – Reproduction of Figure 3

To reproduce the main result of the paper:

bash run_experiment.sh

Or manually:

python scripts/extract_metrics.py
python scripts/run_drl.py
python scripts/plot_figure3.py

---

Expected Output

After execution, the following files should be generated:

data/processed/processed_results.csv
results/baseline/results.csv
results/drl/results.csv
figures/figure3.png

---

Expected Results

The generated figure should present:

Baseline (no control)

• High variability
• Frequent oscillations
• Lower average performance

GlobalAmBC-DRL

• Reduced variability
• More stable behavior
• Higher average success rate

---

Security Considerations

This artifact does not pose any security risks:

• No external network access
• No privileged system operations
• No destructive actions

---

Documentation

Detailed reproduction steps are available in:

docs/reproduction_steps.md

---

Notes

• The dataset figure3_data.csv is provided to ensure deterministic reproduction of the published results.
• The DRL module is included to demonstrate the pipeline structure and system behavior.
• The artifact reproduces the qualitative behavior of the original experiments without requiring OMNeT++ or full DRL training.

---

Final Remarks

This artifact was designed to ensure clarity, simplicity, and reproducibility, enabling reviewers to validate the main claims of the paper with minimal effort.

---

Citation

If you use this artifact, please cite the associated paper (original title in Portuguese):

@inproceedings{galhardo2026globalambc,
  title={Um Novo Módulo Arquitetural de Controle Centralizado com DRL para Redes IoT Densas sem Bateria Baseadas em Retroespalhamento Ambiente},
  author={Galhardo, Edwardes Amaro and Oliveira Junior, Antonio Carlos and Westphall, Carlos Becker},
  booktitle={Simpósio Brasileiro de Redes de Computadores e Sistemas Distribuídos (SBRC)},
  year={2026}
}

You may also cite this artifact as:

@article{galhardo2026artifact,
  title={GlobalAmBC-DRL: A Reproducible Artifact for Dense Batteryless IoT Networks with Ambient Backscatter},
  author={Galhardo, Edwardes Amaro},
  journal={Software/Artifact},
  year={2026},
  url={https://github.com/LABORA-INF-UFG/GlobalAmBC-DRL-SBRC2026},
  note={Research artifact, publicly available repository}
}

---

License

This project is licensed under the MIT License.