The potential for intelligent, AI-controlled robots to 'lend a hand' to humans in hospitals, in elderly and childcare, in factories, in restaurants, in the service industry and in the home is enormous. However, such an interactive robot is only effective if it can react flexibly to different environments and situations. Accordingly, it must be capable of learning. The 13 partners in the EU IntelliMan project have dedicated themselves to this goal.

Over the last few decades, research in the field of intelligent, interactive and adaptive robotics has gained considerable momentum. Robotic arms and robot grippers for machines that can interact directly and autonomously with their environment are becoming increasingly available and affordable. Of central importance in the future, however, will be how such autonomous systems can learn new skills, as the real world has too many variations for an exact model of human desires and behavior, the environment, the objects in it or the skills required to handle them to be created in advance.

The next generation of robotic manipulation systems should be able to use artificial intelligence to autonomously perform tasks with limited human supervision and interact with objects regardless of their material, size and shape. By interacting with humans and their environment, these systems should simultaneously learn new skills that enable them to be used for unforeseen, non-preprogrammed tasks. These applications range from industrial production and logistics to service robotics and wearable devices such as exoskeletons and prostheses. This means that these robotic systems must learn and adhere to certain safety requirements - that is, they must recognize when they cannot perform a task under the applicable safety requirements so that a 'trust level' can be established between humans and robots.

"The IntelliMan project focuses on the question of how a robot can learn efficiently in order to perform tasks in a targeted, high-performance and safe manner," says Prof. Gianluca Palli, coordinator of the IntelliMan project at the Department of Electrical, Electronic and Information Engineering 'Guglielmo Marconi' at the Università di Bologna (Italy). "The robot should learn interaction skills with people, objects and its environment and develop abstractable interaction scenarios from this. The robot should also be able to derive properties and functionalities of the objects through targeted interaction. User perception and acceptance should be taken into account throughout."

The IntelliMan scientists want to investigate problems of gripping and placing flexible objects in complex and different application scenarios: in upper arm prosthetics, in everyday kitchen activities, in the flexible production of cable sets for automotive wiring harnesses and in the handling of fresh food for supermarket logistics applications. This heterogeneous set of usage scenarios with a wide variety of requirements and challenges should make it possible to develop different solutions for the interaction problem.

The IntelliMan consortium consists of scientific and industrial partners from Germany, Italy, Spain, Switzerland, Slovenia and the United Kingdom. The 13 partners come from the fields of AI, robotics, information and communication technologies (ICT) as well as social sciences, humanities and economics and are working together on an interdisciplinary basis in the project under the coordination of the Università di Bologna (Italy). The EU is funding the project under the European Framework Program for Research and Innovation 'Horizon Europe' with 4.5 million euros.

From Bavaria, the German Aerospace Center, Friedrich-Alexander-Universität Erlangen-Nürnberg and the Bavarian Research Alliance (BayFOR) are participating in IntelliMan. They are being funded with a total of around 1.2 million euros.