User-Centric Object Navigation: A Benchmark with Integrated User Habits for Personalized Embodied Object Search
Hongcheng Wang, Jinyu Zhu, Hao Dong
AI summary
Problem
Existing object navigation benchmarks assume objects follow common-sense spatial rules, ignoring individual user placement habits. This limitation reduces the adaptability and real-world effectiveness of navigation agents in personalized home environments.
Approach
The authors introduce UcON, a benchmark that modifies household scenes using over 22,000 personalized user habits, and propose a Habit Retrieval Module to extract relevant habits for LLM-based navigation planning.
Key results
- First large-scale benchmark formalizing habit-conditioned object navigation across 489 categories
- Current SOTA methods show significant performance degradation under habit-driven placements
- Integrating user habits consistently improves navigation success rates
- Proposed Habit Retrieval Module further enhances LLM reasoning and navigation efficiency
Why it matters
It provides a critical evaluation framework for developing home service robots that can adapt to individual human behaviors rather than relying on generic spatial assumptions.
Abstract
In the evolving field of robotics, the challenge of Object Navigation (ON) in household environments has attracted significant interest. Existing ON benchmarks typically place objects in locations guided by general scene priors, without accounting for the specific placement habits of individ- ual users. This omission limits the adaptability of navigation agents in personalized household environments. To address this, we introduce User-centric Object Navigation (UcON), a new benchmark that incorporates user-specific object placement habits, referred to as user habits. This benchmark requires agents to leverage these user habits for more informed decision- making during navigation. UcON encompasses approximately 22,600 user habits across 489 object categories. UcON is, to our knowledge, the first benchmark that explicitly formalizes and evaluates habit-conditioned object navigation at scale and covers the widest range of target object categories. Additionally, we propose a habit retrieval module to extract and utilize habits related to target objects, enabling agents to infer their likely locations more effectively. Experimental results demon- strate that current SOTA methods exhibit substantial perfor- mance degradation under habit-driven object placement, while integrating user habits consistently improves success rates. Code is available at https://github.com/whcpumpkin/ User-Centric-Object-Navigation.