GeNIE: A Generalizable Navigation System for In-the-Wild Environments, Jiaming Wang★, Diwen Liu★, Jizhuo Chen★, Jiaxuan Da, Nuowen Qian, Tram Minh Man, Harold Soh★, arXiv preprint
Links:
Reliable navigation in unstructured, real-world environments remains a significant challenge for embodied agents, especially when operating across diverse terrains, weather conditions, and sensor configurations. We introduce GeNIE (Generalizable Navigation System for In-the-Wild Environments), a robust navigation framework designed for global deployment that sets a new benchmark for outdoor robot navigation.
Key Achievements
🏆 First Place in the Earth Rover Challenge (ERC) at ICRA 2025
- 79% of maximum possible score - outperforming second-best team by 17%
- Zero human interventions throughout the entire competition
- Evaluated across six countries spanning three continents
Technical Contribution
1. SAM-TP: Generalizable Traversability Prediction
- Built on SAM2 architecture for robust terrain understanding
- Enables pixel-wise traversability prediction across diverse environments
- Generalizes across different terrains, weather conditions, and sensor configurations
2. Novel Path Fusion Strategy
- Enhances planning stability in noisy and ambiguous settings
- Provides robust decision-making for autonomous navigation
- Handles uncertainty in real-world outdoor environments
Open Source Contribution
We are committed to advancing the field by releasing:
- 🔧 Complete codebase implementation
- 🎯 Pre-trained model weights
- 📊 Newly curated datasets for real-world navigation research
The SAM-TP repository provides the perception core of GeNIE, enabling researchers to build upon our traversability prediction capabilities.
Resources
You can find our paper here. Check out our SAM-TP repository on GitHub for the perception components.
Citation
Please consider citing our paper if you build upon our results and ideas.
Jiaming Wang★, Diwen Liu★, Jizhuo Chen★, Jiaxuan Da, Nuowen Qian, Tram Minh Man, Harold Soh★, “GeNIE: A Generalizable Navigation System for In-the-Wild Environments”, arXiv preprint
@article{wang2024genie, title={GeNIE: A Generalizable Navigation System for In-the-Wild Environments}, author={Wang, Jiaming and Liu, Diwen and Chen, Jizhuo and Da, Jiaxuan and Qian, Nuowen and Man, Tram Minh and Soh, Harold}, journal={arXiv preprint arXiv:2506.17960}, year={2024} }
Contact
If you have questions or comments, please contact Jiaming or Harold.
Acknowledgements
This research is supported by the National Research Foundation, Singapore under its Medium Sized Center for Advanced Robotics Technology Innovation.
