# 触觉好奇心驱动机器人交互

> 原文链接：https://robotics.hellobike.com/papers/2026-09-30-2609-40134
> 频道：papers | 更新日期：2026-09-30 | 标签：arxiv, cs.RO
> 核心摘要：针对RL操作探索样本低效，TacEx将触觉纳入不确定性好奇心，在抓取、取放和VLA后训练中验证。

# 触觉好奇心驱动机器人交互

::: v-pre
> 原文链接：[Tactile Curiosity Drives Robot Interaction](https://arxiv.org/abs/2609.40134)（arXiv 2609.40134 · 发布日期：2026-09-30）

**Tactile Curiosity Drives Robot Interaction**

针对RL操作探索样本低效，TacEx将触觉纳入不确定性好奇心，在抓取、取放和VLA后训练中验证。

## 核心要点

- 随机探索多在自由空间耗费预算，难以发现产生操作技能的接触。
- TacEx按感知模态分解模型不确定性，将好奇心导向触觉通道。
- 无需任务奖励或专家示范探索，即可发现复杂接触动力学并学习操作、抓取。
- 触觉探索数据支持离线学习取放策略，并用于高效后训练VLA模型。

**验证范围与局限**：验证范围为操作、抓取、取放策略及VLA后训练，摘要未说明仿真或真机。

## 英文摘要 (Abstract)

Mastering robot manipulation skills via reinforcement learning (RL) remains largely sample-inefficient. The most common RL algorithms rely on random action sampling to discover new strategies, resulting in agents that allocate most of their training budget to motions in free space, away from the contacts from which manipulation skills emerge. Existing intrinsic motivation methods based on model disagreement or epistemic uncertainty improve on isotropic noise, but they can also reward uncertainty in functionally irrelevant transitions, such as erratic motions in free space. In this work, we argue that tactile feedback provides a natural signal for exploration, and introduce TacEx, a framework that incorporates touch into epistemic uncertainty-driven exploration by decomposing model uncertainty across sensory modalities and directing curiosity toward the tactile channel. By anchoring curiosity to the sense of touch, TacEx drives the robot to discover complex contact dynamics, learning to manipulate and grasp objects without task rewards or expert demonstrations during exploration. The interaction-dense dataset collected through this tactile-driven curiosity supports offline learning of downstream pick-and-place policies without additional environment interaction. We further use tactile-driven exploration to post-train vision-language-action (VLA) models. Although the VLAs are initially pre-trained without tactile feedback, post-training with TacEx substantially improves downstream performance while remaining highly sample-efficient.

作者：Klemens Iten, Alexander Proshkin, Bhavya Sukhija, Stelian Coros, Andreas Krause, Pieter Abbeel, Carmelo Sferrazza
:::

## 原始参考来源 (Sources)

- [arXiv 2609.40134：Tactile Curiosity Drives Robot Interaction](https://arxiv.org/abs/2609.40134)