Real-world variation
Furniture layouts, object clutter and people sharing the space introduce variation. Fragile objects, deformable materials and multistep chores require adaptable handling.
Back to all scenarios
DATA FOR REAL-WORLD WORK / 01
Home scenarios
Everyday spaces. Extraordinary learning.
Build domestic robot skills with shared data for tidying, folding, cleaning and everyday deliveries.
Sample access depends on the selected dataset. Use a published sample link when available, or request a sample for evaluation.
The scenario
Everyday spaces. Extraordinary learning.
Bring robot learning into living rooms, bedrooms, kitchens and bathrooms. Home scenarios document the complete process from perceiving a scene and understanding an instruction to carrying out household chores. Tasks range from organizing scattered belongings and tidying surfaces to handling tableware and delivering everyday items.
Operating environments
Living rooms, bedrooms, kitchens and bathrooms, with varied furniture arrangements, cluttered surfaces and narrow passageways.
Tasks to learn
- 01Organize belongings
- 02Tidy surfaces
- 03Fold clothes
- 04Handle tableware
- 05Clean floors
- 06Deliver everyday items
Data and signals
Records may include visual, depth and tactile information, joint movements, applied forces and language instructions, linking scene perception to action execution.
Capabilities to develop
Train and evaluate domestic service robots in object handling, deformable material manipulation, task sequencing and shared-space operation.
See the work in context.
Dedicated spaces for scene photography and task videos. Published media will show the environment and the actions captured.
Scenario image
Task demonstration video
Home
Buy more data
Sample access depends on the selected dataset. Use a published sample link when available, or request a sample for evaluation.