Learning-Based Spectrum Cartography in Low Earth Orbit Satellite Networks: An Overview

This repository contains the source code for learning-based spectrum cartography in LEO satellite networks. The associated paper surveys learning-based spectrum cartography for LEO satellite networks, with a focus on attention mechanisms as principled operators for adaptive and reliability-aware measurement fusion across localization, radio map reconstruction, and resource allocation tasks. It reviews the modeling foundations and key challenges of representative tasks, and analyzes how attention-based learning enables flexible fusion of heterogeneous measurements for both inference and map-informed decision-making, supported by representative formulations and simulation studies.

---

Installation

The following software and libraries are required:

• Python 3.11
• PyTorch 2.5.1
• CVXPY
• NVIDIA Sionna RT

Repository Structure

1. example_nw_estimator.py
   Example: NW estimator for reliability-aware LEO satellite localization.

2. example_attention_leo_localization.py
   Example: Attention-based fusion for LEO satellite localization.

3. case_study_attention_gps_leo.py
   Case study: Attention-based GPS correction in LEO-assisted localization.

4. example_radioMapRecon_with_sionna.py
   Example: Reliability-aware radio map reconstruction.

5. models_and_analysis_attention_rem.py
   Model and analysis: Radio map reconstruction via learnable attention.

6. starlink_sionna.py
   Example: TLE-driven Starlink satellite selection and Sionna-based LEO ground-scene visualization for satellite localization analysis.

7. water-level.py
   Example: Water-filling based joint transmit and interference power allocation analysis for multi-channel reliability optimization.

8. radio_map_color_occlusion.py
   Example: Occlusion-aware radio map reconstruction and data acquisition visualization in urban sensing environments.

9. leo_nw_attention.py
   Model and analysis: NW-attention smoothing for robust LEO localization trajectory recovery under noisy and outlier-contaminated measurements.

10. leo_case_study.py
    Case study: CRLB-based geometry analysis for LEO localization, showing how LOS diversity improves geometry-aware satellite fusion over SNR-only selection.

---

Usage

Running the Python scripts

Scripts such as example_attention_leo_localization.py can be executed in two ways:

1. Direct Execution

Open the file in a Python IDE (e.g., PyCharm) and run it directly.

2. Command-Line Execution

From a terminal, navigate to the script directory and run:

python example_attention_leo_localization.py