*4.6. Human in the Loop and Other Sociotechnical Concerns*

In many application environments, robots interact with humans. For example, humans may use the services of a robot team. Alternatively, humans may augment the capabilities of the team, effectively making them a member of the team, or humans may give instructions to a robot team to control the execution. Research in social robotics delivered various means of interacting with robots, e.g., based on voice or gestures. However, interaction with and control of a whole team of robots received little attention so far. Some general questions remain. How would the team and the individual team members be addressed? How would an "emergency button" be implemented that immediately stops the execution of the whole MRT? How would the possible actions of a human be modeled in the team plan? How would the MRT react to unanticipated actions of the human?

Sociotechnical concerns for technical innovations, as presented in Section 3.8 above, received increasing attention in society recently. Clearly, these concerns also apply to robot teamwork. As in other technical domains, the big question is as follows: how do we translate abstract sociotechnical requirements that are mostly specified in natural language into concrete engineering artefacts for multi-robot applications? For example, liability issues for a team of heterogeneous robots from different manufacturers could be difficult to decide. Likewise, explanations for team decisions and actions may be even more difficult to understand for a user who demands transparency for MRT activities; this will undermine the user´s trust in the technology and may lead to a lack of acceptance. Experts from different disciplines must work together to solve these interdisciplinary puzzles.
