The AI-PRISM consortium holds the necessary expertise to strategically improve the EU Smart manufacturing value chain in four domains: collaborative environment, robot perception and recognition, robot programming and human factors. We are working to achieve the right trade-off between complementary and overlapping among the industrial partners.
In the industrial category, we count on use case partners with experience in actual market product development and industrial activities such as KEBA. Together with the other four use cases, we generate a broad range of user-centred requirements, deployment blueprints, integration plans, testing scenarios, and performance monitoring to assess preliminary designs and feed the next iterations.
KEBA’s pilot site covers the discrete manufacturing sector in Austria. KEBA is an internationally active technology company headquartered in Linz/Austria and has 26 subsidiaries in 16 countries. They have been developing and producing pioneering automation solutions for many industries under the maxim “Automation by innovation” for over 50 years. In this use case, KEBA’s initial AI-based Automation Control System prototype will be enhanced to enable users’ natural interaction with robots, including training and execution phases of human-robot collaboration.
Today we have KEBA’s team sharing their role in the AI control for natural, multi-modal collaboration. More concretely, the customer site application will be developed in the AI-PRISM project. The focus is on one or more control systems and multiple channels for communication. The overall goals of this application are flexibility and speed in the training and execution phases.
KEBA aims to be “easy to use,” with solutions oriented to the needs of the user that create the optimum connection to technical problem-solving. Ultimately, their technologies help people to make the world of life and work easier. Therefore, at AI-PRISM, together with research partner Profactor (PRO), they want to achieve a more natural interaction between humans and robots without programming.
The initial prototype provides an architecture with interfaces to ROS and relevant machine learning libraries, an initial technical basis to improve the speed and flexibility of (robotic) control on the shop floor. However, the solution needs more support for the multi-modal training of robots for complex and collaborative shop floor processes. The challenge here is that the robotic system (incl. gripers) is variable (from the control system point of view).
With an overall robotic-process training and execution cycle, the AI-PRISM modules placed on AI-Control of KEBA will allow them to work with minimal end-user training most naturally, teaching by example. Furthermore, multi-modal, ambient sensor networks will help to determine deviations in the execution to determine the need for adaptation. It might include, for example, the sound of machines in case of wrong configurations or broken tools. Ultimately, KEBA expects to decrease mental load and time during operations and training, decrease setup time and technical knowledge of users needed for operation and training and increase the number of communication modes.
Do you want to learn more about AI-PRISM and the team behind the project? Subscribe to our newsletter and follow us on LinkedIn and Twitter, so you don’t miss a thing!