A concrete and clear definition of the term social robotics is difficult - as is often the case in robotics. Therefore, a variety of definitions exist, for example according to Graaf et al., with an identification of eight essential factors for social robots in the private environment: mutual interaction, showing thoughts and feelings, social awareness, social support, autonomy, comfort, similarity to oneself and mutual respect. According to Sarrica et al., many attempts at a definition also have in common that robots are social if they autonomously perceive environmental stimuli, react to them, interact with humans or other robots and can also understand and follow general social rules. In addition to the attempts at a definition, it can be noted that a large number of publications on the topic of social robotics have so far primarily outlined systems in a medical/nursing context.

Extending the envisaged application scenarios for social robots to the industrial sector would make perfect sense. For example, large numbers of autonomous transport robots have long been sharing logistics and transport areas in production environments with very different groups of people. When, shortly after a shift change, a large number of jammed transport robots cacophonously ask to clear the way in front of a temporary maintenance barrier, one quickly wishes for a better adapted behavior beyond the manual configuration of fleet managers. If the robots were able to better classify the actions of people in terms of time, and if they took into account that the 20th repetition of the same request would most likely also remain unanswered, systems with better autonomous adaptation capabilities could be implemented overall. The fact that an advanced animal-like service robot could technically perform complex tasks completely autonomously, but requires a human companion for real-life applications in order to intercept the unpredictable reactions of other people, also highlights the need for the development of social skills. Only a social service robot can autonomously reassure anxious people in appropriate situations by explaining its own functionality and intentions and at the same time calmly but firmly ask people who start to playfully test the system without being asked to do so to "let it work in peace".

Regulatory issues

In addition to numerous technical challenges in the development of social service robots for industrial use, there are also regulatory issues that must be taken into account. For example, it must be clarified whether the foreseeable ban on AI-supported recording and classification of emotions in the work context in the EU's AI Act also applies to the implementation of functions to increase the acceptance of service robots. In general, systems must be developed that can analyze situations with local data processing without storing personal data.

From January 2024, a new research network of the Bavarian Research Foundation entitled FORSocialRobots will address the challenges in the development and utilization of social robots for various applications and explicitly also for industrial applications. In this network, many of the relevant Bavarian players from industry, application and research will work together over three years to research important issues in social robotics and transfer findings into a wide variety of applications. The Chair of FAPS at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) is coordinating the sub-project Architecture of Social Skills in order to develop ROS-based software solutions for educable (industrial) robots. The Fraunhofer Institute IGCV's Processing Technology division and the Fraunhofer Institute IIS's Smart Sensing and Electronics division are responsible for the social situational communication sub-project. The sub-project socially adaptive and proactive interaction is being carried out at the Chair of HCAI at the University of Augsburg. The sub-project Simulation and Validation of Socially Cognitive Robots in the Digital Twin is based at the Chair of PI, also at the University of Augsburg. Finally, another sub-project, Human-Robot Interaction in the Work Context, is being carried out at the Chair of PiA at FAU. Together, they are working on the design of goal-oriented and accepted holistic social communication and cooperation between robots and humans in order to increase the acceptance of robots in the scenarios mentioned as examples.