Soft-Bodied Intelligence for Manipulation

SoMa

SOMA-Logo
SOMA-Logo
Credit:
PR

SoMa – Soft-Bodied Intelligence for Manipulation – aims to develop resilient, robust, easy-to-program and therefore cost-efficient hands for manipulation and secure gripping. Constraints of the environment are not understood as an obstacle, but as an opportunity. For example, when slicing onions, humans use the front surface of the hand as a guide to stabilize the knife. We want to learn from such examples in SoMa and to test and verify the soft manipulation prototypes developed in such areas as the food industry.

  
Runtime:
2015-05-01 until 2019-04-30
Project partners:
• Technische Universität Berlin (Coordinator), TUB, Germany
• University of Pisa, UNIPI, Italy
• Italian Institute of Technology, IIT, Italy
• German Aerospace Center, DLR, Germany
• Institute of Science and Technology Austria, IST, Austria
• Ocado Ltd., OCADO, United Kingdom
• Disney Research, DISNEY, Switzerland
Website:
Fields of application:
• Food and Agriculture Industry
• Online Supermarket
• human/robot coexisting and co-working environments
• Soft Manipulation
Funding
 

Project details

The biggest obstacle to the widespread use of advanced robotic systems for manipulating random industrial objects is the complexity, susceptibility, low force and difficulty of handling the robotic systems. The EU project SoMa seeks to resolve this conflict, leading to the development of simple, yielding, sufficiently strong, robust and easy-to-program robotic systems for manipulation and secure gripping. The idea is soft manipulation (SoMa).

Constraints of the environment – SoMa explores a new direction for robotic manipulation with the environment, as opposed to manipulation of or in it. In this approach, the physical constraints imposed on each object by its environment and the gripper itself are not considered obstacles. On the contrary, SoMa understands the constraints as an opportunity to pre-adjust compliant grippers to the shape of the object and the environment, thus simplifying gripping and manipulation. The analysis of the resulting possibilities will enable robust gripping and manipulation in dynamic and rapidly changing environments.

Soft Manipulation – The key to exploiting environmental constraints is robotic hands that are able to yield; that have an inherent ability to adapt to features of the environment. The traditional approach to robotic manipulation is in direct opposition to SoMa: State-of-the-art gripper designs are focused on rigid hands and objects, trying to find algorithmic solutions to inherently complex, often under-constrained problems. This means that our approach to gripper design for manipulating and gripping objects in unknown environments goes far beyond the state of the art. Nonetheless, humans show identical patterns of gripping and manipulation, which motivates us to fundamentally change robotic manipulation.

In SoMa, we want to develop resilient robotic hands for versatile and reliable use of environmental constraints and the associated robust, cost-effective and secure gripping designs.

The developed soft manipulation technology will be used in an industrial context of the agricultural and food industry. This aims at handling and manipulating irregularly shaped, yielding and delicate goods, such as fruits and vegetables. We will test and verify the developed soft manipulation grippers in this industrial environment, thus dramatically improving robotic gripping and manipulation.