Alongside the contract of their choice, SPRIND will support the Polybot team with intensive coaching as well as strategic development advice as well as the ability to connect with investors. The Polybot team will present their concept to nearly 300 investors who are interested in attending the Venture SPRIND meeting in Berlin the month of April in 2025.

“This contract demonstrates that Polybot isn’t just an idea from the ivory tower rather a practical technology for sustainable, future-proof agricultural practices,” claims project manager Wieland Brendel of the Max Planck Institute for Intelligent Systems as well as the ELLIS Institute in Tubingen, Germany.

A close relationship between Tubingen AI Center and SPRIND represents a huge achievement for the Tubingen AI community. Polybot is an excellent demonstration that shows how the Tubingen AI Center links research with practical applications. “This initial validation from outside is a huge boost for our team,” says Martin Kiefel Technical Director of Polybot. “With this validation it is now possible to build our algorithm for learning to tackle the most difficult tasks of farming, such as harvesting fine vegetables. We can also verify them alongside farmers on the farm.”

Bernhard Scholkopf, Scientific Director of the ELLIS Institute and Director at the Max Planck Institute for Intelligent Systems Adds: “Excellent research unfolds its potential when it doesn’t just creates knowledge, but assists in solving the problems that we face today.”

A more sustainable approach to agriculture

Polybot is an entirely automated solution for cultivating fruit, crops, and other vegetables with the latest AI technology. Polybot aims at automating various tasks, such as pulling tomatoes, weeding or cucumbers, or cutting them free of charge. With the help of computer vision as well as robot mechanical systems, Polybot helps reduce the use of chemical herbicides. Additionally, it encourages small-scale, sustainable agriculture.

Automating manual labor tasks allows farmers to utilize their labor efficiently, and could increase the yield as time passes. The heart of the system is an autonomous machine that is equipped with a precise and efficient manipulator capable of carrying out complicated tasks such as tomato harvesting. Its control system is based on an innovative machine-learning pipeline that allows the robot to quickly learn new tasks through farmer demonstrations–eliminating the need for time-consuming programming.

Agriculture is undergoing innovation

The present validation project will test the effectiveness of Polybot for harvesting fine vegetable–a project that requires precision, as well as sophisticated 3D vision. It is a difficult but a suitable test area to test the robot’s capability to perform tasks which have previously been deemed to be economically impossible.

In addition to testing on technical aspects The project is also focused on the potential benefits to farmers. The harvesting of fine vegetables is mostly manual which is causing labor shortages and ongoing problems. With the help of farmers, the initiative aims to identify the requirements needed to create a product on the market that can provide value.

“The Polybot project leverages recent developments in machine-learning to make the practice of polyculture farming financially feasible. We are thrilled to have the opportunity to bring such concepts into reality at the Tubingen AI center – concepts which are beneficial to society and only possible by cutting-edge research” states Matthias Bethge, Director of the Tubingen AI Center.

More information is available here:

Linda Behringer

Max Planck Institute for Intelligent Systems

Tel: +49 151 2300 1111

Email: [email protected]

www.mpg.de