A Behavioral Approach to Visual Navigation
with Graph Localization Networks

Kevin Chen, Juan Pablo de Vicente, Gabriel Sepúlveda, Fei Xia, Alvaro Soto, Marynel Vázquez, Silvio Savarese


main figure

Overview

Inspired by research in psychology, we introduce a behavioral approach for visual navigation using topological maps. Our goal is to enable a robot to navigate from one location to another, relying only on its visual input and the topological map of the environment. We propose using graph neural networks for localizing the agent in the map, and decompose the action space into primitive behaviors implemented as convolutional or recurrent neural networks. Using the Gibson simulator, we verify that our approach outperforms relevant baselines and is able to navigate in both seen and unseen environments.


Results on Seen Areas

Trajectory Length: ~39.82m
Trajectory Length: ~38.44m
Trajectory Length: ~28.59m

Results on Unseen Areas

Trajectory Length: ~12.28m
Trajectory Length: ~15.23m
Trajectory Length: ~18.24m

Links

If you find our project helpful, please consider citing us:

@article{chen2019graphnav,
  title={A Behavioral Approach to Visual Navigation with Graph Localization Networks},
  author={Chen, Kevin and de Vicente, Juan Pablo and Sepulveda, Gabriel and Xia, Fei and Soto, Alvaro and Vazquez, Marynel and Savarese, Silvio},
  journal={arXiv preprint arXiv:1903.00445},
  year={2019}
}
        

Video Summary


Acknowledgements

Toyota Research Institute ("TRI") provided funds to assist the authors with their research but this article solely reflects the opinions and conclusions of its authors and not TRI or any other Toyota entity. This work is also partially funded by Fondecyt grant 1181739, Conicyt, Chile.