Update README.md
Browse filesUpdate README.md to add detailed task descriptions
README.md
CHANGED
|
@@ -19,6 +19,29 @@ This dataset was created using [LeRobot](https://github.com/huggingface/lerobot)
|
|
| 19 |
- **Paper:** [More Information Needed]
|
| 20 |
- **License:** apache-2.0
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
## Dataset Structure
|
| 23 |
|
| 24 |
[meta/info.json](meta/info.json):
|
|
|
|
| 19 |
- **Paper:** [More Information Needed]
|
| 20 |
- **License:** apache-2.0
|
| 21 |
|
| 22 |
+
- **Task Objective:** Take the water bottle and simulate pouring water into a transparent glass cup.
|
| 23 |
+
- **Operational Objects:** Transparent glass cup (7cm diameter × 8cm height) + 350ml Wahaha bottled water (Purchase link: 【淘宝】https://e.tb.cn/h.60Y4A3cvIpzq2xO?tk=iMs8eJrmMWW CZ001 「娃哈哈 纯净水饮用水350ml*24瓶整箱瓶装企业会议用水」点击链接直接打开 或者 淘宝搜索直接打开).
|
| 24 |
+
- **Operation Duration:** Each operation takes approximately 20 seconds.
|
| 25 |
+
- **Recording Frequency:** 30 Hz.
|
| 26 |
+
- **Robot Type:** 7-DOF dual-arm G1 robot.
|
| 27 |
+
- **End Effector:** Three-fingered dexterous hands.
|
| 28 |
+
- **Dual-Arm Operation:** Yes.
|
| 29 |
+
- **Image Resolution:** 640x480.
|
| 30 |
+
- **Camera Positions:** Wrist-mounted (monocular camera) + head-mounted (binocular cameras).
|
| 31 |
+
- **Data Content:**
|
| 32 |
+
- • Robot's current state.
|
| 33 |
+
- • Robot's next action.
|
| 34 |
+
- • Current camera view images.
|
| 35 |
+
- **Robot Initial Posture:** The first robot state in each dataset entry.
|
| 36 |
+
- **Object Placement:** Place the water cup adjacent to the three-fingered dexterous hand on the robot's left arm in its initial position (arm vertical); position the water bottle within reach of the right hand.
|
| 37 |
+
- **Camera View:** Follow the guidelines in **Part 5** of [AVP Teleoperation Documentation](https://github.com/unitreerobotics/avp_teleoperate).
|
| 38 |
+
|
| 39 |
+
- **Important Notes:**
|
| 40 |
+
1. Due to the inability to precisely describe spatial positions, adjust the scene to closely match the first frame of the dataset after installing the hardware as specified in **Part 5** of [AVP Teleoperation Documentation](https://github.com/unitreerobotics/avp_teleoperate).
|
| 41 |
+
2. Data collection is not completed in a single session, and variations between data entries exist. Ensure these variations are accounted for during model training.
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
|
| 45 |
## Dataset Structure
|
| 46 |
|
| 47 |
[meta/info.json](meta/info.json):
|