引导、思考、行动：视觉-语言-动作模型中的交互式具身推理

Yiran Ling, Qing Lian, Jinghang Li, Qing Jiang, Tianming Zhang, Xiaoke Jiang, Chuanxiu Liu, Jie Liu, Lei Zhang

2026

摘要

本文提出GTA-VLA框架，通过允许用户以显式视觉线索引导机器人策略，实现空间可控的具身推理。该框架将外部空间引导与内部任务规划统一为空间-视觉思维链，解决了现有模型在域外偏移和错误纠正方面的局限性。

vision-language-actionembodied reasoninghuman-robot interactionspatial guidancechain-of-thought

HRI

📊 3 引用

Kai Ding, Qingyuan Mao, Yaqian Zhang 等 6 位作者

Robotics and Computer-Integrated Manufacturing · 2026

HRI

📊 0 引用

Yingchao You, Ze Ji, Changyun Wei

Robotics and Computer-Integrated Manufacturing · 2026

HRI

📊 0 引用

Jiahui Si, Wenchao Li, Xi Chen 等 7 位作者

Robotics and Computer-Integrated Manufacturing · 2026

HRI

📊 0 引用

Rui Qian, Xi Zhang, Dongpeng Li 等 5 位作者

Robotics and Computer-Integrated Manufacturing · 2026