While a robot can easily pick up a steel plate, grabbing a towel is notoriously difficult—fabrics wrinkle, droop, and tangle, changing shape by the second, which renders traditional computer vision completely useless. Based in Munich, Germany, sewts solves this exact problem by combining machine learning with image processing to help robots understand the state of soft materials and predict how they will deform the moment they are lifted.
Features & Use Cases
The core technology centers on state estimation and grasp planning for deformable objects. Its most mature commercial application is in industrial laundries, where hotels and hospitals process massive volumes of textiles. Separating tangled piles of towels and bedsheets, flattening them, and feeding them into ironers has historically been a heavily labor-dependent process.
It is ideal for industrial laundries, apparel manufacturing automation, and logistics sorting. Just like in global markets, hospitality and healthcare laundry services are labor-intensive and face severe labor shortages. Automation technology specifically designed for "soft goods" is far better suited to real-world demands than standard, general-purpose robotic arms.
Key Features
- Soft material state visual recognition
- Deformation prediction and grasp planning
- Industrial laundry automation applications
- Integration with robotic arm systems
- Specialized AI models for textile processing
Common Use Cases
- Industrial laundry textile sorting
- Garment manufacturing automation
- Textile quality inspection
- Flexible material loading and unloading
Key Features
- Soft material state visual recognition
- Deformation prediction and grasp planning
- Industrial laundry automation applications
- Integration with robotic arm systems
- Specialized AI models for textile processing
Pros
- Solves the notoriously difficult problem of handling deformable objects in robotics
- Clear, highly practical real-world deployment in industrial laundries
- Directly targets industries suffering from severe labor shortages
Cons
- Application scope remains relatively focused
- High initial investment for the complete solution (including robotic arms)
- Implementation requires modifications to on-site workflows
Use Cases
- Industrial laundry textile sorting
- Garment manufacturing automation
- Textile quality inspection
- Flexible material loading and unloading
Editor's Note
There is a well-known saying in robotics: the things humans find easiest are the hardest for machines. A three-year-old can fold a towel, yet the world's top robotics teams have been working on it for over a decade and are still perfecting it.
FAQ
Why is picking up fabric so difficult for robots?
Because fabric lacks a fixed shape. Rigid objects only require position and orientation data to be picked up, but fabric changes shape the exact moment it is lifted. The system must accurately predict the deformation outcome in order to plan the next step.
Can garment sewing be fully automated?
Not yet. Sewing involves complex two-handed coordination and tension control. Current automation successes are primarily in the front and back ends (sorting, flattening, and folding), while intermediate sewing processes still heavily rely on manual labor.