Update README.md
Browse files
README.md
CHANGED
|
@@ -2,8 +2,24 @@
|
|
| 2 |
license: mit
|
| 3 |
---
|
| 4 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
MODAL : Data Confuser "Small files only, Big files confusion"
|
| 6 |
|
| 7 |
CHARS : 517253f76137x0p
|
| 8 |
|
| 9 |
-
DCVERSION : 0.0.1.2 built 260210
|
|
|
|
| 2 |
license: mit
|
| 3 |
---
|
| 4 |
|
| 5 |
+
|
| 6 |
+
# MaNI Series
|
| 7 |
+
```
|
| 8 |
+
The MaNI series is a real-world simulation dataset for robotic manipulation, focusing on complex
|
| 9 |
+
human-object interaction and fine-grained motion modeling. This dataset systematically covers various manipulation
|
| 10 |
+
types common in real-world robotic operation scenarios, including grasping, pushing, pulling, rotating, assembly,
|
| 11 |
+
and tool use. It models real-world contact, friction, constraints, and dynamics through high-fidelity physical simulation.
|
| 12 |
+
MaNI employs a multi-modal data acquisition and synchronization mechanism, including RGB/Depth visual observations,
|
| 13 |
+
6-DoF poses of objects and end-effectors, joint states, force/torque feedback, and continuous motion trajectories,
|
| 14 |
+
providing rich and stable data support for learning end-to-end manipulation strategies from perception to control.
|
| 15 |
+
Unlike purely simulated or purely real-world datasets, MaNI reduces the sim-to-real gap by reconstructing real-world
|
| 16 |
+
scenarios and calibrating parameters in the simulation environment. This dataset is particularly suitable for research
|
| 17 |
+
directions such as imitation learning, reinforcement learning, and manipulation planning, and can serve as an important
|
| 18 |
+
benchmark for robotic skill learning, generalization ability evaluation, and complex interaction reasoning.
|
| 19 |
+
```
|
| 20 |
+
|
| 21 |
MODAL : Data Confuser "Small files only, Big files confusion"
|
| 22 |
|
| 23 |
CHARS : 517253f76137x0p
|
| 24 |
|
| 25 |
+
DCVERSION : 0.0.1.2 built 260210
|