WorldFly: A World-Model-Based Vision-Language-Action Model for UAV Navigation

2026-06-04

AI Summary

This work addresses the challenge of robust end-to-end drone navigation in dense urban environments, where severe occlusions and abrupt viewpoint changes lead to incomplete observations and hinder reliable decision-making. To overcome this, the authors propose WorldFly, a novel framework that integrates a world model into a vision-language-action system for aerial navigation. WorldFly employs a dual-branch coupled flow-matching mechanism to jointly generate future video predictions and navigation actions, thereby enhancing spatial understanding and enabling imagination-driven policy planning. The study also introduces the Urban Canyon Traversal Benchmark for systematic evaluation. Experimental results demonstrate that WorldFly significantly outperforms existing baselines on this benchmark, exhibiting superior robustness and generalization—particularly in unseen urban scenarios.

Abstract

End-to-end Vision-Language-Action (VLA) models have shown promise in UAV navigation. However, existing approaches typically rely on historical observations to directly predict actions, often struggling in dense urban environments where severe occlusions and sharp turns result in drastic viewpoint transitions. We argue that the ability to "imagine" future states -- inherent in World Models -- is critical for robust decision-making under such partial observability. To address this, we construct a challenging Urban Canyon Traversal Benchmark, specifically designed to evaluate spatial understanding in scenarios characterized by severe occlusions and drastic viewpoint transitions. To this end, we propose WorldFly, a novel world-model-based VLA framework that employs a dual-branch coupled flow matching mechanism to jointly generate future video predictions and navigation actions, thereby explicitly guiding the agent's policy via spatial imagination. Extensive evaluations on our benchmark demonstrate that WorldFly outperforms other baselines, particularly in unseen environments, validating the effectiveness of integrating world models into embodied aerial agents.

Problem

• UAV navigation
• occlusions
• viewpoint transitions
• partial observability
• spatial understanding

Innovation

• World Model
• Vision-Language-Action
• Future Prediction
• UAV Navigation
• Spatial Imagination