Robotic systems represent a crucial component of tomorrow's medicine and have the potential to sustainably address current challenges such as nursing shortages and increasing workloads. Despite available technologies, their application has yet to be fully realized, due to the specific requirements of the healthcare sector. Inpatient care, for example, is a highly dynamic and extremely complex field characterized by stringent safety, hygiene, and ethical considerations. Unlike many industrial applications, robotic systems in this sector must interact directly with people and integrate seamlessly into human-led processes.
Objective:
Within the KIRA-med project, a voice-controlled robotic assistant is being developed for routine tasks on patient wards. The robotic assistance focuses on direct physical interaction with nursing staff during patient-related activities. The robot's behavior can be dynamically adapted to the requirements of various clinical routine situations via natural language interaction.
Scientific questions:
What does intuitive and efficient natural language control of assistive robots look like in the context of inpatient care (language, amount of information, timing, feedback, context understanding)?
Can the integration of LLM-supported speech interaction unlock the existing potential of assistive robotics for inpatient care, and does this lead to a quantifiable reduction in the physical and cognitive workload of healthcare professionals during routine care?
How can the robot's behavior be made transparent to users and allow for immediate correction at any time?
