# DexTOG框架发布：语言引导扩散模型实现灵巧手任务导向抓取

> 原文链接：https://robotics.hellobike.com/papers/2025-04-06-2504-04573-7f3601b7
> 频道：papers | 更新日期：2025-04-06 | 标签：auto, arxiv, cs.RO
> 核心摘要：提出语言引导扩散学习框架DexTOG及8万条数据集DexTOG-80K，解决灵巧手高自由度空间下任务导向抓取的非唯一最优解难题，确立多任务灵巧操作新基准。

# DexTOG框架发布：语言引导扩散模型实现灵巧手任务导向抓取

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[DexTOG: Learning Task-Oriented Dexterous Grasp with Language](https://arxiv.org/abs/2504.04573)（arXiv 2504.04573）

## 核心内容

针对灵巧手任务导向抓取（TOG）中高自由度配置空间搜索难、最优姿态非唯一等痛点，研究提出语言引导扩散学习框架DexTOG。该工作构建了包含80个物体、5大类别的DexTOG-80K数据集，基于Shadow Robot Hand验证了多任务灵巧操作能力，为具身智能灵巧操作提供了新的算法范式与数据基准。

技术上，DexTOG集成扩散生成模型DexDiffu与配套数据引擎，利用自然语言指令约束生成过程，在复杂任务约束下精准规划多模态抓取姿态。相比传统双指夹爪方法，该框架有效应对灵巧手多关节耦合挑战，支持特定任务下的多重有效抓取解搜索，并通过仿真验证了其在非结构化场景中的泛化性能与鲁棒性。

## 关键看点

- **算法架构**：采用语言引导扩散模型DexDiffu，结合数据引擎解决高DOF空间中任务导向抓取的非唯一最优解生成问题
- **数据资产**：发布DexTOG-80K数据集，涵盖5类80种物体，基于Shadow Hand采集以支撑多任务灵巧抓取训练
- **核心突破**：克服传统双指夹爪局限，实现语言语义约束下的高自由度灵巧手多模态抓取姿态规划与仿真验证

## 工程与产业观察

该工作将扩散模型引入灵巧手任务导向抓取，通过语言条件约束解决了高维动作空间的多解歧义问题。配套的DexTOG-80K数据集填补了语义驱动灵巧操作数据空白，为具身智能从简单拾取向复杂工具使用与精细作业演进提供了关键的算法与数据基础设施。

## 英文摘要 (Abstract)

This study introduces a novel language-guided diffusion-based learning framework, DexTOG, aimed at advancing the field of task-oriented grasping (TOG) with dexterous hands. Unlike existing methods that mainly focus on 2-finger grippers, this research addresses the complexities of dexterous manipulation, where the system must identify non-unique optimal grasp poses under specific task constraints, cater to multiple valid grasps, and search in a high degree-of-freedom configuration space in grasp planning. The proposed DexTOG includes a diffusion-based grasp pose generation model, DexDiffu, and a data engine to support the DexDiffu. By leveraging DexTOG, we also proposed a new dataset, DexTOG-80K, which was developed using a shadow robot hand to perform various tasks on 80 objects from 5 categories, showcasing the dexterity and multi-tasking capabilities of the robotic hand. This research not only presents a significant leap in dexterous TOG but also provides a comprehensive dataset and simulation validation, setting a new benchmark in robotic manipulation research.

作者：Jieyi Zhang, Wenqiang Xu, Zhenjun Yu, Pengfei Xie, Tutian Tang, Cewu Lu
:::

## 原始参考来源 (Sources)

- [arXiv 2504.04573：DexTOG: Learning Task-Oriented Dexterous Grasp with Language](https://arxiv.org/abs/2504.04573)