NavWareSet SFM Analysis

Overview

This repository contains data and analysis tools for studying robot–human interactions using the NavWareSet dataset.
It focuses on calibrating and evaluating the Social Force Model (SFM) with selected robot–pedestrian interaction tracks.

Contents

• NavWareSet_SFM.ipynb
  Jupyter notebook implementing:

  • Data loading from selected track files
  • Preprocessing and synchronization
  • Simulation of pedestrian trajectories using the Social Force Model
  • RMSE-based evaluation of model fit
  • Visualization of real vs. simulated trajectories

• CSV track files (s_track_sceneXX_colY_from_<timestamp>.csv)
  Extracted subsets of the NavWareSet dataset. Each file corresponds to a specific scene and collection column.

  • sceneXX → scenario ID (e.g., scene 21, 34, 47, 8)
  • colY → collection column (e.g., different pedestrians or robot tracks)
  • from_<timestamp> → dataset export time or unique identifier

• README.md
  Repository documentation.

Getting Started

Requirements

• Python 3.9+
• Jupyter Notebook or Google Colab
• Core libraries: numpy, pandas, matplotlib, scipy
• UAIbotPy (SFM simulation support)

Usage

1. Open the notebook locally with Jupyter, or click the Colab badge above.
2. Upload the CSV track files if running in Colab.
3. Run the notebook cells step by step to reproduce the simulations and analysis.

Example Workflow

• Select a track file from s_track_scene*.csv
• Run the SFM simulation with initial parameters
• Optimize parameters with L-BFGS-B
• Compare simulated trajectories with real pedestrian data
• Save and visualize results as plots or animations

License

This repository is released under the MIT License. See LICENSE for details.

Citation

If you use this code or data in your research, please cite the NavWareSet dataset paper:

Brayan, J., Deng, S., Alves Neto, A., Okunevich, I., Krajnik, T., Bremond, F., & Yan, Z.
NavWareSet: A Dataset of Socially Compliant and Non-Compliant Robot Navigation.