Source: https://actionable.ai/datasets/U01.html

[Skip to content](https://actionable.ai/datasets/U01.html#main)

MANIPULATION / UMI / AA U01

# Bimanual UMI

Connect human demonstrations to precise gripper movement and action.

[Enquire about this dataset ↗](https://actionable.ai/?dataset=U01#contact)[Browse all datasets ↗](https://actionable.ai/#catalogue)

COLLECTION INVENTORY10,000+ episodes

SAMPLE RECORDINGS10

Sample access, licensing, and delivery are confirmed for each enquiry.

CAPTURE DETAILS

## Sample specifications

| Capture system | Two handheld grippers + head-mounted camera

| Resolution | 3 × 1600 × 1200

| Frame rate | 60 fps

| Format | H.264 · MP4 + MCAP

| Views | 3 video streams

| Views | Head, left gripper and right gripper

| Pose and action | 6DoF gripper pose and opening; paired state/action at 60 Hz

| IMU | 60 Hz, aligned per frame

| Camera alignment | Reported three-camera offset: 22–40 ms

| Lens model | KannalaBrandt8

| Calibration | Independent intrinsics and extrinsics for each camera

INSIDE THE DATASET

## Package and coverage

Three video views, pose/action sequences, camera calibration and a trajectory-projection preview per episode.

### Annotation

Task-level labels; can be reorganized to a customer-defined skill taxonomy.

### Coverage

Household and office manipulation across ten sample tasks.

REPRESENTATIVE WORK

## Sample tasks

- Placing fruit in a bowl
- Organizing condiment bottles
- Clearing dishes
- Setting tableware
- Sorting laundry
- Collecting toys
- Storing household items
- Organizing computer peripherals
- Sorting snacks
- Folding clothes

EXPLORE THE CATALOGUE

## Find the right data for your model.

[Next: Real Robot · G1 / G2 ↗](https://actionable.ai/datasets/R01.html)[All eight collections ↗](https://actionable.ai/#catalogue)