Search papers, labs, and topics across Lattice.
This paper introduces DaViNCi, the first outdoor Vision-and-Language Navigation (VLN) dataset that incorporates continuous actions and dynamic elements, addressing the limitations of existing datasets that rely on fixed discrete topological graphs. By evaluating agents on six distinct maps with 6,933 trajectories, the authors reveal a significant drop in success rates鈥攐ver 10% in discrete environments and even more in continuous settings鈥攈ighlighting the increased complexity of real-world navigation. The findings underscore the importance of adapting VLN frameworks to better reflect unpredictable outdoor scenarios, thereby enhancing the sim-to-real transfer for navigation agents.
Success rates plummet by over 10% when navigating dynamic outdoor environments, revealing the inadequacy of traditional VLN datasets.
Vision-and-Language Navigation (VLN) has progressively expanded from indoor to outdoor environments. However, existing outdoor VLN datasets still rely on fixed discrete topological graphs for construction. It fails to align with the rapidly changing real-world outdoor environments and impedes the sim-to-real transfer of VLN agents. To address this limitation, we propose DaViNCi (\textbf{D}yn\textbf{a}mic \textbf{Vi}sion-and-Language \textbf{N}avigation in \textbf{C}ont\textbf{i}nuous Environment), the first outdoor VLN dataset that simultaneously introduces both continuous and dynamic factors. The agent not only moves in the outdoor environment using continuous actions but is also required to handle unpredictable dynamic elements. The dataset encompasses six distinct maps with a total of 6,933 trajectories. Through comprehensive comparative experiments, we find that the success rate on DaViNCi decreased by more than 10\% in discrete environments compared to previous datasets. And there is an even greater decline in continuous settings, demonstrating the challenge of DaViNCi. Furthermore, we clarify the impact of action granularity and dynamic elements. These results demonstrate the practical value of DaViNCi in advancing outdoor VLN toward more realistic environments. The website is https://xzh0312.github.io/DaViNCi/.
