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

WorldFly: A World-Model-Based Vision-Language-Action Model for UAV Navigation arXiv:2606.06147v1 Announce Type: new 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 -- inhere