aiqbench analysis
NovaPlan advances robotic autonomy by integrating Vision-Language Models with video generation to enable zero-shot planning for complex, long-horizon manipulation tasks. The system addresses critical reliability issues like physical hallucinations through a hierarchical framework that verifies visual plans and employs hybrid geometric representations to ensure geometric consistency during execution. This approach allows robots to perform multi-stage assembly and recover from errors without requiring task-specific training or demonstrations.
NovaPlan: Zero-Shot Long-Horizon Manipulation via Closed-Loop Video Language Planning
arXiv:2602.20119v2 Announce Type: replace-cross Abstract: Solving complex long-horizon robotic tasks requires joint reasoning over abstract task structure and low-level physical interaction. While combining Vision-Language Models (VLMs) and video generation models offers a promising path for zero-shot planning, their individual tendencies to hallucinate physics or violate geometric consistency often compound over time, preventing reliable real-world execution. We introduce NovaPlan, a hierarchical framework that enables robust, zero-shot long-horizon manipulation by systematically proposing, verifying, and repairing visual plans. At the high level, a VLM planner decomposes tasks and filters out dynamically inconsistent futures by verifying multiple candidate video rollouts. To translate these imagined futures into reliable physical actions, NovaPlan utilizes a hybrid geometric representation that adaptively switches between object-centric flow and human hand flow. Finally, NovaPlan closes the loop by continuously monitoring execution to verify outcomes and synthesize local, non-prehensile corrective behaviors, such as fingertip poking, when failures occur. Across diverse multi-stage tasks, NovaPlan substantially outperforms prior zero-shot systems, achieving complex assembly and dexterous error recovery entirely without task-specific training or demonstrations. Please visit our project website for additional results: https://nova-plan.github.io/
Abstract:Solving complex long-horizon robotic tasks requires joint reasoning over abstract task structure and low-level physical interaction. While combining Vision-Language Models (VLMs) and video generation models offers a promising path for zero-shot planning, their individual tendencies to hallucinate physics or violate geometric consistency often compound over time, preventing reliable real-world execution. We introduce NovaPlan, a hierarchical framework that enables robust, zero-shot long-horizon manipulation by systematically proposing, verifying, and repairing visual plans. At the high level, a VLM planner decomposes tasks and filters out dynamically inconsistent futures by verifying multiple candidate video rollouts. To translate these imagined futures into reliable physical actions, NovaPlan utilizes a hybrid geometric representation that adaptively switches between object-centric flow and human hand flow. Finally, NovaPlan closes the loop by continuously monitoring execution to verify outcomes and synthesize local, non-prehensile corrective behaviors, such as fingertip poking, when failures occur. Across diverse multi-stage tasks, NovaPlan substantially outperforms prior zero-shot systems, achieving complex assembly and dexterous error recovery entirely without task-specific training or demonstrations. Please visit our project website for additional results: this https URL
Accepted to CoRL 2026. Project webpage: this https URL
Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2602.20119 [cs.RO]
(or arXiv:2602.20119v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2602.20119 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 284 words
Related coverage
SkillWeaver: Agentic Exploration over Neural Interaction Skills for Scalable Robot Data Generation
arXiv · Sep 30, 2026
Embodied-R1.5: Evolving Physical Intelligence via Embodied Foundation Models
arXiv · Jun 12, 2026
Exploratory, Communicative, and Deployable: Vision-Driven Embodied Agents for Open-World Mobile Manipulation
arXiv · Jul 28, 2026
[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model
Latent Space · Oct 13, 2026