Understanding and navigating human-centered environments over extended periods of time while considering human behavior and routines remains a fundamental challenge in robotics. In real-world settings, robots may be asked to locate a specific individual, predict where that person is likely to be, or estimate when they typically leave a building. Addressing such queries requires reasoning over extensive histories of observations and capturing long-term behavioral patterns.
To this end, we introduce Human-Centered Memory for Embodied Robots (HUMEMBR), a system designed for embodied question answering and routine-conditioned navigation. HUMEMBR integrates a continuous memory construction process with a parallel retrieval and querying mechanism, enabling the system to accumulate structured representations of human routines while supporting interactive, user-driven queries.
Our experimental results indicate that HUMEMBR improves long-horizon reasoning about human behavior relative to full-context LLM baselines, while using substantially fewer tokens. Furthermore, we deploy \papername on a physical robot in two distinct environments, showing its ability to handle diverse queries and navigation tasks under real-world conditions.
@inproceedings{huber2026humembr,
title = {HUMEMBR: Learning Human Routines for Predictive Embodied Navigation},
author = {Huber, Samira and Pelzer, Klaas and Nguyen, Duc M. and Xiao, Xuesu and Pirk, S{\"o}ren},
booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
year = {2026},
}