Abstract

Object-goal navigation (ObjectNav) in multi-floor scenarios presents a challenge due to sparse rewards caused by long-horizon decision-making. In this paper, we propose a diagnostic study based on a modular framework with an effective learnable policy to analyze failure factors in multi-floor scenarios. To achieve an effective policy for diagnosis, we design the hierarchical factorization policy that deconstructs a single global policy into an intra-floor exploration policy and an inter-floor switching policy. To providing an effective initialization for Reinforcement Learning (RL), the lightweight intra-floor policy is learned by distilling the exploration logic of Visual Language Models (VLMs). Under idealized assumptions, we show that the factorized policy is theoretically equivalent to a single global policy at the policy-representation level. Experiment results indicate that perception performance and stair climbing stability are the primary bottlenecks in multi-floor navigation.

Keywords

Subject

Publication details

Journal
Not available
Open access
Green open access

Cite this article

APA 7

Zhai, S., Ye, S., Xiong, R., & Wang, Y. (2026). Learning Modular Policy for Multi-Floor Object Navigation:A Factorized Framework for Diagnostic Study. https://omanscience.com/en/articles/learning-modular-policy-for-multi-floor-object-navigation-a-factorized-framework-for-diagnostic-study

MLA 9

Zhai, Shichao, et al. "Learning Modular Policy for Multi-Floor Object Navigation:A Factorized Framework for Diagnostic Study." https://omanscience.com/en/articles/learning-modular-policy-for-multi-floor-object-navigation-a-factorized-framework-for-diagnostic-study.

Chicago (author–date)

Zhai, Shichao, Shuhao Ye, Rong Xiong, and Yue Wang. 2026. "Learning Modular Policy for Multi-Floor Object Navigation:A Factorized Framework for Diagnostic Study." https://omanscience.com/en/articles/learning-modular-policy-for-multi-floor-object-navigation-a-factorized-framework-for-diagnostic-study.

Harvard

Zhai, S., Ye, S., Xiong, R. and Wang, Y. (2026) 'Learning Modular Policy for Multi-Floor Object Navigation:A Factorized Framework for Diagnostic Study', Available at: https://omanscience.com/en/articles/learning-modular-policy-for-multi-floor-object-navigation-a-factorized-framework-for-diagnostic-study.

Vancouver

Zhai S, Ye S, Xiong R, Wang Y. Learning Modular Policy for Multi-Floor Object Navigation:A Factorized Framework for Diagnostic Study. https://omanscience.com/en/articles/learning-modular-policy-for-multi-floor-object-navigation-a-factorized-framework-for-diagnostic-study

IEEE

S. Zhai, S. Ye, R. Xiong, and Y. Wang, "Learning Modular Policy for Multi-Floor Object Navigation:A Factorized Framework for Diagnostic Study," https://omanscience.com/en/articles/learning-modular-policy-for-multi-floor-object-navigation-a-factorized-framework-for-diagnostic-study.