Search papers, labs, and topics across Lattice.
This paper introduces the Uncertainty-Aware Navigation World Model (UA-NWM) for aerial image-goal navigation, addressing the limitations of existing methods that rely on a few point predictions in uncertain outdoor environments. By formulating trajectory scoring as conditional out-of-distribution detection, UA-NWM effectively captures plausible futures through an uncertainty subspace, allowing for robust scoring based on unexplainable prediction discrepancies. Extensive experiments and real-world UAV tests show that UA-NWM outperforms current models while ensuring low inference latency, making it a significant advancement in aerial navigation technology.
Trajectory scoring in aerial navigation can be revolutionized by focusing on unexplainable prediction discrepancies, leading to more robust and efficient UAV navigation.
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
Tsinghua AI