Hierarchical Object-Oriented POMDP Planning for Object Rearrangement

ICLR 2026 Conference Submission22421 Authors

Keywords: rearrangement, POMDP, planning, reinforcement learning, object search

TL;DR: We propose a Hierarchical Object-Oriented POMDP planning approach for multi-object rearrangement in partially observable environments. We show it's effectiveness in the AI2Thor simulator experiments

Abstract: We present an online planning approach and a new benchmark dataset for solving multi-object rearrangement problems in partially observable, multi-room environments. Current object rearrangement solutions, primarily based on Reinforcement Learning or hand-coded planning methods, often lack adaptability to diverse challenges. To address this limitation, we propose a Hierarchical Object-Oriented Partially Observed Markov Decision Process (HOO-POMDP) planner that leverages object-factored belief representations for efficient multi-object rearrangement. This approach comprises of (a) an object-oriented POMDP planner generating sub-goals, (b) a set of low-level policies for sub-goal achievement, and (c) an abstraction system converting the continuous low-level world into a representation suitable for abstract planning. To enable rigorous evaluation of rearrangement challenges, we introduce MultiRoomR, a comprehensive benchmark featuring diverse multi-room environments with varying degrees of partial observability (10-30\% initial visibility), blocked paths, obstructed goals, and multiple objects (10-20) distributed across 2-4 rooms. Experiments demonstrate that our system effectively handles these complex scenarios while maintaining robust performance even with imperfect perception, achieving promising results across both existing benchmarks and our new MultiRoomR dataset.

Primary Area: applications to robotics, autonomy, planning

Submission Number: 22421