Human-like Navigation in a World Built for Humans

Bhargav Chandaka, Gloria Xinyue Wang, Haozhe Chen, Henry Che, Albert J. Zhai, Shenlong Wang
Proceedings of The 9th Conference on Robot Learning, PMLR 305:1790-1808, 2025.

Abstract

When navigating in a man-made environment they haven’t visited before—like an office building—humans employ behaviors such as reading signs and asking others for directions. These behaviors help humans reach their destinations efficiently by reducing the need to search through large areas. Existing robot navigation systems lack the ability to execute such behaviors and are thus highly inefficient at navigating within large environments. We present ReasonNav, a modular navigation system which integrates these human-like navigation skills by leveraging the reasoning capabilites of a vision-language model (VLM). We design compact input and output abstractions based on navigation landmarks, allowing the VLM to focus on language understanding and reasoning. We evaluate ReasonNav on real and simulated navigation tasks and show that the agent successfully employs higher-order reasoning to navigate efficiently in large, complex buildings.

Cite this Paper

BibTeX
@InProceedings{pmlr-v305-chandaka25a, title = {Human-like Navigation in a World Built for Humans}, author = {Chandaka, Bhargav and Wang, Gloria Xinyue and Chen, Haozhe and Che, Henry and Zhai, Albert J. and Wang, Shenlong}, booktitle = {Proceedings of The 9th Conference on Robot Learning}, pages = {1790--1808}, year = {2025}, editor = {Lim, Joseph and Song, Shuran and Park, Hae-Won}, volume = {305}, series = {Proceedings of Machine Learning Research}, month = {27--30 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v305/main/assets/chandaka25a/chandaka25a.pdf}, url = {https://proceedings.mlr.press/v305/chandaka25a.html}, abstract = {When navigating in a man-made environment they haven’t visited before—like an office building—humans employ behaviors such as reading signs and asking others for directions. These behaviors help humans reach their destinations efficiently by reducing the need to search through large areas. Existing robot navigation systems lack the ability to execute such behaviors and are thus highly inefficient at navigating within large environments. We present ReasonNav, a modular navigation system which integrates these human-like navigation skills by leveraging the reasoning capabilites of a vision-language model (VLM). We design compact input and output abstractions based on navigation landmarks, allowing the VLM to focus on language understanding and reasoning. We evaluate ReasonNav on real and simulated navigation tasks and show that the agent successfully employs higher-order reasoning to navigate efficiently in large, complex buildings.} }
Endnote
%0 Conference Paper %T Human-like Navigation in a World Built for Humans %A Bhargav Chandaka %A Gloria Xinyue Wang %A Haozhe Chen %A Henry Che %A Albert J. Zhai %A Shenlong Wang %B Proceedings of The 9th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2025 %E Joseph Lim %E Shuran Song %E Hae-Won Park %F pmlr-v305-chandaka25a %I PMLR %P 1790--1808 %U https://proceedings.mlr.press/v305/chandaka25a.html %V 305 %X When navigating in a man-made environment they haven’t visited before—like an office building—humans employ behaviors such as reading signs and asking others for directions. These behaviors help humans reach their destinations efficiently by reducing the need to search through large areas. Existing robot navigation systems lack the ability to execute such behaviors and are thus highly inefficient at navigating within large environments. We present ReasonNav, a modular navigation system which integrates these human-like navigation skills by leveraging the reasoning capabilites of a vision-language model (VLM). We design compact input and output abstractions based on navigation landmarks, allowing the VLM to focus on language understanding and reasoning. We evaluate ReasonNav on real and simulated navigation tasks and show that the agent successfully employs higher-order reasoning to navigate efficiently in large, complex buildings.
APA
Chandaka, B., Wang, G.X., Chen, H., Che, H., Zhai, A.J. & Wang, S.. (2025). Human-like Navigation in a World Built for Humans. Proceedings of The 9th Conference on Robot Learning, in Proceedings of Machine Learning Research 305:1790-1808 Available from https://proceedings.mlr.press/v305/chandaka25a.html.

Related Material