用於空中影像目標導航的不確定性感知世界模型

摘要

空中影像目標導航要求無人機（UAV）抵達由目標影像指定的目標位置。現有的基於世界模型的方法利用預測的未來狀態對候選軌跡進行排序，但通常僅依賴於一個或少數幾個點預測，這在具有大量未來狀態不確定性的大規模戶外環境中並不足夠。為了解決此限制，我們提出了不確定性感知導航世界模型（UA-NWM），這是一種用於空中影像目標導航的高效潛在世界模型，其將軌跡評分表述為條件式分佈外偵測。UA-NWM 以不確定性子空間表示可能的未來狀態，並將預測與目標之間的差異分解為可歸因於不確定性的成分與無法解釋的成分。僅使用無法解釋的殘差進行評分，從而無需多個未來樣本即可實現穩健的選擇。大量實驗表明，UA-NWM 在保持低推論延遲的同時，持續優於現有的導航世界模型。真實世界的無人機實驗進一步驗證了其實際應用性。專案頁面：https://duryi.github.io/UA-NWM-Project-Page

English

Aerial image-goal navigation requires an unmanned aerial vehicle (UAV) to reach a target location specified by a goal image. Existing world-model-based methods rank candidate trajectories using predicted futures, but typically rely on only one or a few point predictions, which is inadequate for large-scale outdoor environments with substantial future-state uncertainty. To address this limitation, we propose the Uncertainty-Aware Navigation World Model (UA-NWM), an efficient latent world model for aerial image-goal navigation, which formulates trajectory scoring as conditional out-of-distribution detection. UA-NWM represents plausible futures with an uncertainty subspace and decomposes the prediction--goal discrepancy into uncertainty-explainable and unexplainable components. Only the unexplainable residual is used for scoring, enabling robust selection without multiple future samples. Extensive experiments demonstrate that UA-NWM consistently outperforms existing navigation world models while maintaining low inference latency. Real-world UAV experiments further validate its practical applicability. Project page: https://duryi.github.io/UA-NWM-Project-Page