andreaskoepf/bottle_test

---
license: apache-2.0
task_categories:
- robotics
tags:
- LeRobot
- RoboTwin
- bimanual
- simulation
configs:
- config_name: default
data_files: data/*/*.parquet
---

This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).

<a href="https://huggingface.co/spaces/lerobot/visualize_dataset?path=andreaskoepf/robotwin_dk1_realcam_stack_blocks_two">
<img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl.svg"/>
<img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl-dark.svg"/>
</a>

## Dataset Description

Bimanual manipulation demonstrations for the **Stack Blocks Two** task, generated in the [RoboTwin](https://github.com/TianxingChen/RoboTwin) simulator with domain randomization.

- **Robot:** TRLC DK1 bimanual (`bi_dk1_follower`) - 2x 6-DOF arms with parallel grippers
- **Task:** Stack Blocks Two
- **Episodes:** 100
- **Total frames:** 21,752 (21.3 min @ 17 fps)
- **Avg episode length:** 12.8s

### Cameras

| Camera | Resolution | Codec | FPS |
|---|---|---|---|
| `observation.images.head` | 640x360 | h264 | 17 |
| `observation.images.left_wrist` | 640x360 | h264 | 17 |
| `observation.images.right_wrist` | 640x360 | h264 | 17 |

### Action & State Space

- **`action`**: float32[14] - joint position targets (6 joints + 1 gripper per arm)
- **`observation.state`**: float32[14] - current joint positions + gripper state

### Task Descriptions

- "the right arm grabs red block, places it in the center, the left arm stacks green block on red block."
- "Pick up red block and green block, set them at the center, then position green block above red block."
- "the left arm grabs red block, places it in the center, the right arm stacks green block on red block."
- "Position red block at the center and stack green block above it."
- "Put red block in the center, then add green block above it."

## Loading the Dataset

```python
from lerobot.datasets.lerobot_dataset import LeRobotDataset

dataset = LeRobotDataset("andreaskoepf/robotwin_dk1_realcam_stack_blocks_two")
frame = dataset[0]
print(frame["observation.state"].shape) # torch.Size([14])
print(frame["action"].shape) # torch.Size([14])
```

## Citation

```bibtex
@misc{robotwin2025,
title = {RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins},
author = {Chen, Yao Mu et al.},
year = {2025},
url = {https://github.com/TianxingChen/RoboTwin}
}
```

---
许可证：Apache-2.0
任务类别：
- 机器人学
标签：
- LeRobot
- RoboTwin
- 双臂操作
- 仿真
配置项：
- 配置名称：默认
数据文件：data/*/*.parquet
---

本数据集基于[LeRobot](https://github.com/huggingface/lerobot)项目构建。

<a href="https://huggingface.co/spaces/lerobot/visualize_dataset?path=andreaskoepf/robotwin_dk1_realcam_stack_blocks_two">
<img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl.svg"/>
<img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl-dark.svg"/>
</a>

## 数据集概述

本数据集为**堆叠积木二（Stack Blocks Two）**任务的双臂操作演示数据，基于[RoboTwin](https://github.com/TianxingChen/RoboTwin)仿真平台生成，并采用了领域随机化技术。

- **机器人平台**：TRLC DK1双臂机器人（`bi_dk1_follower`）——配备2台6自由度机械臂与平行夹爪
- **任务名称**：堆叠积木二（Stack Blocks Two）
- **演示轨迹数**：100
- **总帧数**：21752（以17 FPS采集，总时长约21.3分钟）
- **单轨迹平均时长**：12.8秒

### 相机配置

| 相机视角 | 分辨率 | 编码格式 | 帧率 |
|---|---|---|---|
| `observation.images.head` | 640×360 | H.264 | 17 |
| `observation.images.left_wrist` | 640×360 | H.264 | 17 |
| `observation.images.right_wrist` | 640×360 | H.264 | 17 |

### 动作与状态空间

- **`action`（动作空间）**：float32[14] 类型，对应关节位置目标（每台机械臂含6个关节与1个夹爪，总计14维）
- **`observation.state`（观测状态空间）**：float32[14] 类型，包含当前关节位置与夹爪状态

### 任务指令描述

以下为该任务的自然语言指令示例：
1. 右臂抓取红色积木并放置于中心位置，左臂将绿色积木堆叠至红色积木之上。
2. 拾取红色积木与绿色积木并放置于中心区域，随后将绿色积木堆叠于红色积木上方。
3. 左臂抓取红色积木并放置于中心位置，右臂将绿色积木堆叠至红色积木之上。
4. 将红色积木放置于中心区域，并将绿色积木堆叠于其上方。
5. 将红色积木放置于中心位置，随后在其上方添加绿色积木。

## 数据集加载

python
from lerobot.datasets.lerobot_dataset import LeRobotDataset

dataset = LeRobotDataset("andreaskoepf/robotwin_dk1_realcam_stack_blocks_two")
frame = dataset[0]
print(frame["observation.state"].shape) # torch.Size([14])
print(frame["action"].shape) # torch.Size([14])

## 引用格式

bibtex
@misc{robotwin2025,
title = {RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins},
author = {Chen, Yao Mu et al.},
year = {2025},
url = {https://github.com/TianxingChen/RoboTwin}
}