You presented the new Soto robot at Logimat. How does it differ from the previous Toru picking robot?
Brantner: Soto can handle significantly larger and heavier objects than Toru. This is particularly important when it comes to industrial applications. In contrast to an e-commerce warehouse, small load carriers up to a size of 600 x 400 x 400 millimetres and a weight of 15 kilograms often have to be moved here. Soto is perfectly equipped for this task - but the basic functions are the same as those of Toru.

Do you have any further innovations planned in the near future?
Brantner: We are now primarily working on making the Toru and Soto robots even more robust for continuous use. We are also currently evaluating whether and how we will make the Acros (Advanced Cooperative Robot Operating System) software stack, which turns robots into intelligent and self-acting machines, available to other robot manufacturers as a product.

In the last financing round, you gained new strategic investors such as Zalando, the technology group Körber and Fiege Logistik. How will you use the capital increase?
Brantner: Primarily to drive forward the technical development of the robots and to expand the sales structure in Germany and abroad.

Magazino has been growing steadily since it was founded in 2014, and robotics and human-robot collaboration are also booming. Do you plan to conquer other sectors as well?
Brantner: As a start-up, it's a good idea to focus. We are concentrating on intralogistics processes in e-commerce and the manufacturing industry. We believe that these areas have huge potential.

More than two thirds of your employees are computer scientists. What role does computer science play in robotics in particular?
Brantner: The biggest! The robots can perceive their environment using hardware such as 3D cameras and numerous different sensors. But the interpretation of the environment and the subsequent decision-making - that all takes place in the software. Especially when robots are operating in the same area as humans, they have to be able to find their way in an environment that is constantly changing.

What do you think the future of logistics will look like? Will we soon be able to manage without people?
Brantner: Certainly not. With their motor skills and talent for improvisation, humans have a huge advantage over robots. However, flexible automation solutions such as Toru or Soto will see massive growth over the next few years. Robots will not be able to do everything, but they will be able to automate certain activities in an economically viable and flexible way. For example, robots can take over tasks that are ergonomically unfavorable for humans: reaching from the lowest or highest shelf level, orders with long walking distances or working hours at night or on public holidays. It is therefore much more important that the robots can work in parallel with humans. Then both do what they do best.

How is it that picking robots like Soto or Toru are developed by a start-up, but not by one of the big robot manufacturers?
Brantner: The challenges for robots like ours are extremely complex. They could not be solved with previous automation technology. We therefore needed radically new approaches and methods. However, for large manufacturers who already have existing products and functioning business models, this always means a risk. What's more, this complexity requires a lot of staff with new programming skills. The way in which intelligent robots are programmed today differs fundamentally from previous automation technology. We have recruited many employees in this area directly from the Technical University of Munich or from other universities around the world. With over 40 software developers, we now have one of the largest teams in Europe working on this type of robotics.

What is the difference between perception-controlled robots and conventional industrial robots and what are the challenges?
Brantner: The classic industrial robot works extremely deterministically. A specific movement sequence is programmed into it once and then executed thousands of times. Very repeatable and precise - but without any awareness of the environment or its interpretation. In contrast, perception-controlled robots have the ability to "see" the world around them via 3D cameras and sensors. Using software - and this is the key to intelligent behavior - the robots can then interpret the world, make decisions at runtime and thus adapt their behavior to the respective circumstances.

When working together with humans in the same environment, there will always be deviations from the expected conditions. The big challenge is to provide robots with intelligent algorithms so that they can master situations in continuous operation that we haven't even thought about yet.

Mr. Brantner, thank you very much for the interview!