Newest papers

SPAR: Search-trained Push Action Ranker for Efficient Navigation Among Movable Obstacles

Dhruv Metha Ramesh, Tri-An Nguyen, Junshan Huang, Jingjin Yu, Kostas E. Bekris, Abdeslam Boularias · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28362 · cited by 0 works (OpenAlex) · open-access PDF

French

Pour un robot poussant, le progrès peut dépendre d'une poussée de configuration qui permet une ouverture ultérieure sans le nettoyage d'un passage. Le nombre de séquences candidates peut croître de manière exponentielle avec la longueur de la séquence, ce qui rend la recherche basée sur la physique coûteuse. Nous introduisons le Search-trained Push Action Ranker (SPAR), qui apprend à donner la priorité aux primitifs poussés à partir de résultats de recherche simulés. Les recherches sur les séquences de deux poussées par objet fournissent une surveillance qui distingue les ouvertures immédiates des pousses de configuration. Un transformateur formé par la recherche basée sur la régression et le classement de la recherche scores les primitifs en utilisant des jet

Machine translation (Meta M2M100); the English original follows.

Abstract

Navigation Among Movable Obstacles (NAMO) requires a robot to rearrange objects that block access to its goal. For a pushing robot, progress may depend on a setup push that enables a later opening without itself clearing a passage. The number of candidate sequences can grow exponentially with sequence length, making physics-based search costly. We introduce the Search-trained Push Action Ranker (SPAR), which learns to prioritize push primitives from simulated search outcomes. Searches over sequences of up to two pushes per object provide supervision that distinguishes immediate openings from setup pushes. A transformer trained with regression and ranking losses scores primitives using contact tokens conditioned on scene geometry and reachability regions. The same model guides simulator evaluations in SPAR Search and selects actions for direct execution in SPAR Policy. Both modes solve successive keyhole problems that open access to neighboring free-space regions within a NAMO planner. Across successful runs on $300$ simulated two-push keyhole problems, SPAR Search reduces median simulator calls by $7.0\times$, leading to a $3.8\times$ reduction in planning time relative to Random Search. Without retraining, the same ranker guides search on $600$ keyhole problems which have no solution within two pushes. At a depth limit of five pushes, SPAR Search achieves $50.7\%$ success within $100$ simulator calls, compared with $20.2\%$ for Random Search. These gains carry over to NAMO planning: on hard simulated scenes, SPAR Search achieves $47\%$ success within $100$ simulator calls, compared with $21\%$ for Random Search. In real-robot experiments, SPAR Policy succeeds in $37/40$ trials without a physics-based digital twin, compared with $33/40$ for SPAR Search and $27/40$ for Random Search.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Spar, Computer science, Simulation, Random search, Artificial intelligence, Computer vision

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange