SeqWalker: Sequential-Horizon Vision-and-Language Navigation with Hierarchical Planning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908753262018560 |
|---|---|
| author | Han, Zebin Wang, Xudong Liu, Baichen Lyu, Qi Shang, Zhenduo Dong, Jiahua Liu, Lianqing Han, Zhi |
| author_facet | Han, Zebin Wang, Xudong Liu, Baichen Lyu, Qi Shang, Zhenduo Dong, Jiahua Liu, Lianqing Han, Zhi |
| contents | Sequential-Horizon Vision-and-Language Navigation (SH-VLN) presents a challenging scenario where agents should sequentially execute multi-task navigation guided by complex, long-horizon language instructions. Current vision-and-language navigation models exhibit significant performance degradation with such multi-task instructions, as information overload impairs the agent's ability to attend to observationally relevant details. To address this problem, we propose SeqWalker, a navigation model built on a hierarchical planning framework. Our SeqWalker features: i) A High-Level Planner that dynamically selects global instructions into contextually relevant sub-instructions based on the agent's current visual observations, thus reducing cognitive load; ii) A Low-Level Planner incorporating an Exploration-Verification strategy that leverages the inherent logical structure of instructions for trajectory error correction. To evaluate SH-VLN performance, we also extend the IVLN dataset and establish a new benchmark. Extensive experiments are performed to demonstrate the superiority of the proposed SeqWalker. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_04699 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SeqWalker: Sequential-Horizon Vision-and-Language Navigation with Hierarchical Planning Han, Zebin Wang, Xudong Liu, Baichen Lyu, Qi Shang, Zhenduo Dong, Jiahua Liu, Lianqing Han, Zhi Robotics Artificial Intelligence Sequential-Horizon Vision-and-Language Navigation (SH-VLN) presents a challenging scenario where agents should sequentially execute multi-task navigation guided by complex, long-horizon language instructions. Current vision-and-language navigation models exhibit significant performance degradation with such multi-task instructions, as information overload impairs the agent's ability to attend to observationally relevant details. To address this problem, we propose SeqWalker, a navigation model built on a hierarchical planning framework. Our SeqWalker features: i) A High-Level Planner that dynamically selects global instructions into contextually relevant sub-instructions based on the agent's current visual observations, thus reducing cognitive load; ii) A Low-Level Planner incorporating an Exploration-Verification strategy that leverages the inherent logical structure of instructions for trajectory error correction. To evaluate SH-VLN performance, we also extend the IVLN dataset and establish a new benchmark. Extensive experiments are performed to demonstrate the superiority of the proposed SeqWalker. |
| title | SeqWalker: Sequential-Horizon Vision-and-Language Navigation with Hierarchical Planning |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2601.04699 |