aiqbench analysis
ImagineUAV addresses the brittleness of standard Vision-Language-Action models in UAVs by integrating latent video diffusion to simulate environmental evolution before kinodynamic planning refines collision-free trajectories. This approach achieves real-time execution with only 1.3B parameters, outperforming baselines on both benchmarks and physical flights. The work demonstrates that imagination-driven world modeling is a viable path toward robust, instruction-following aerial navigation.
ImagineUAV: Aerial Vision-Language Navigation via World-Action Modeling and Kinodynamic Planning
arXiv:2606.01205v2 Announce Type: replace Abstract: Vision-language navigation (VLN) for UAVs demands grounding free-form instructions into 6-DoF flight under partial observability. While Vision-Language-Action (VLA) models excel at semantic reasoning, they suffer from brittleness due to geometric inconsistency and dynamics mismatch. To address this, we propose ImagineUAV, an imagination-driven framework leveraging cascaded world-action modeling. Instead of direct regression, ImagineUAV employs a latent video diffusion model to generate instruction-conditioned future observations, explicitly imagining environmental evolution, from which 6-DoF motions are inferred via an action extractor. A kinodynamic planner then refines these estimates into collision-free trajectories. Additionally, a step-distilled inference pipeline ensures real-time execution. With only 1.3B parameters, ImagineUAV outperforms prior VLN and VLA baselines on benchmarks and real-world flights, validating the practicality of imagination-driven aerial navigation.
Abstract:Vision-language navigation (VLN) for UAVs demands grounding free-form instructions into 6-DoF flight under partial observability. While Vision-Language-Action (VLA) models excel at semantic reasoning, they suffer from brittleness due to geometric inconsistency and dynamics mismatch. To address this, we propose ImagineUAV, an imagination-driven framework leveraging cascaded world-action modeling. Instead of direct regression, ImagineUAV employs a latent video diffusion model to generate instruction-conditioned future observations, explicitly imagining environmental evolution, from which 6-DoF motions are inferred via an action extractor. A kinodynamic planner then refines these estimates into collision-free trajectories. Additionally, a step-distilled inference pipeline ensures real-time execution. With only 1.3B parameters, ImagineUAV outperforms prior VLN and VLA baselines on benchmarks and real-world flights, validating the practicality of imagination-driven aerial navigation.
Robotics (cs.RO)
arXiv:2606.01205 [cs.RO]
(or arXiv:2606.01205v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.01205 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 · 209 words
Related coverage
WareFly-VLA: A Vision-Language-Action Framework for UAV Navigation and Human Tracking in Smart Warehouses
arXiv · Oct 7, 2026
WorldFly: A World-Model-Based Vision-Language-Action Model for UAV Navigation
arXiv · Jun 6, 2026
AeroDPO: Unleashing Lightweight UAV Navigation with High-Fidelity Perception and Automated Preference Optimization
arXiv · Aug 27, 2026
Vertical Advantage: Transforming Industries with Lakebase and Agentic AI
Databricks · Aug 27, 2026