Efficient robot navigation inspired by honeybee learning flights

This repository contains all code for the paper:

Efficient robot navigation inspired by honeybee learning flights D. Ou¹, J. J. Hagenaars¹, M. Jankowski¹, M. V. M. Firlefyn¹, C. De Wagter¹, F. T. Muijres², J. Degen³, G. C. H. E. de Croon¹ Nature, 2026. DOI: 10.1038/s41586-026-10461-3

Affiliations:

1. MAVLab, Control and Operations department, Faculty of Aerospace Engineering, Delft University of Technology (TU Delft), the Netherlands.
2. Experimental Zoology Group, Wageningen University, the Netherlands.
3. Navigation Biology Group, Institute of Biology and Environmental Sciences, Carl von Ossietzky University of Oldenburg, Germany.

Repository Structure

Folder	Description	Paper Reference
code_theoretical_simulation_analysis/	Theoretical analysis and ANN simulations	Supplementary Information 1–5
path_integration_noise_simulation/	Path integration noise modelling	Figure 2 a,b
visual_simulator_docker/	Isaac Sim 4.2 visual homing experiments (Docker)	Figures 2 c,d,e
robot_network_training/	Offline network training (laptop and Raspberry Pi)	Figure 3,4
robot_onboard/	Onboard drone code (ROS 2 flight control + camera)	Figure 3,4

Quick Start

Each folder contains its own README.md with setup instructions. In brief:

1. Theoretical analysis — cd code_theoretical_simulation_analysis && pip install -r requirements.txt && python run_all_experiments.py
2. Path integration — cd path_integration_noise_simulation && pip install -r requirements.txt && python run_path_integration_experiments.py
3. Visual simulator — Requires Docker + NVIDIA GPU. See visual_simulator_docker/README.md for setup.
4. Robot network training — cd robot_network_training/home-learning_laptop and follow README.md.
5. Robot onboard — Requires Raspberry Pi 4 + PX4 flight controller. See robot_onboard/README.md.

External Data

• Isaac Sim 3D assets (~2.4 GB): Download from SURFdrive and place in visual_simulator_docker/isaac-sim_assets/.
• Robot training data: Download from the link in robot_network_training/README.md.

Citation

@article{ou2026efficient,
  title   = {Efficient robot navigation inspired by honeybee learning flights},
  author  = {Ou, D. and Hagenaars, J. J. and Jankowski, M. and Firlefyn, M. V. M. and De Wagter, C. and Muijres, F. T. and Degen, J. and de Croon, G. C. H. E.},
  journal = {Nature},
  year    = {2026},
  doi     = {10.1038/s41586-026-10461-3}
}

License

This project is licensed under the MIT License. See LICENSE for details.