ROBO.KIWI

The flagship project ROBO.KIWI is developing new approaches to preserve the experiential knowledge of nursing professionals over the long term and make it available for training purposes. Using wearable sensors, artificial intelligence, and humanoid robotics, nursing practices are recorded, analyzed, and made accessible as interactive learning experiences.

When expertise is lost

Demographic change poses major challenges for the nursing profession. Depending on future trends, Germany could face a shortage of between 280,000 and 690,000 nursing professionals by 2049, while demand continues to rise due to an increasing elderly population. At the same time, rising workloads are leading to more absences and, consequently, intensifying the problem of understaffing.

When experienced nursing staff retire, it is not just human resources that are leaving. Valuable experiential knowledge is also at risk of being lost: motion routines, interaction strategies, and situational decisions developed over years or decades of working experience. This knowledge is often difficult to document or share through traditional teaching methods.

Preserving experiential knowledge for the future of healthcare

This is exactly where ROBO.KIWI comes in. The research project examines how experiential knowledge in nursing can be captured using body-worn sensors, digitally preserved, and made available for training through the use of humanoid robotics. The goal is to develop an innovative learning system that not only demonstrates nursing expertise but also allows students to experience it physically. This enables trainees to benefit from the experience of long-standing nursing professionals, even when those professionals are no longer directly available.

 

Embodied AI for nursing education

ROBO.KIWI combines sensing technologies, artificial intelligence, and humanoid robotics to preserve experiential knowledge and make it available for future generations of learners. Rather than acting as tools, robots become interactive physical learning partners. The project's methodology serves as a blueprint for capturing and transferring experiential knowledge in other domains where expertise is difficult to document and at risk of being lost, such as manufacturing, skilled trades, and agriculture.

Capturing and digitizing know-how

Real-world nursing procedures, such as repositioning and transfers, are captured using the body-worn sensor network maphera® and documented in a multimodal dataset.

Robots as carriers of expertise

The captured motion and interaction data are converted into a data model, curated, and fused into representative reference motions. Subsequently, they are transferred to a humanoid robot system using AI.

Robots learning partners

In training scenarios, the robot plays a dual role: as a nursing professional, it demonstrates the motion sequences of experienced nursing staff; as a care recipient, it simulates realistic reactions and objections.

Methodology: From expertise to an interactive learning partner

The ROBO.KIWI project’s approach combines sensor technology, artificial intelligence, and robotics.

Using the mobile, decentralized maphera® sensor network, physiological signals and motion sequences – optimized over decades – are recorded directly in day-to-day care. The data collected serves as the basis for a multimodal dataset from which AI algorithms identify characteristic motion and interaction patterns.

At the heart of the project is the concept of Embodied AI. The humanoid robot does not simply record knowledge; it makes it physically accessible as an interactive learning medium. As a nursing professional, it demonstrates curated motion sequences, and as a care recipient, it simulates realistic reactions and objections.

The evaluation takes place in a so-called “Skills Lab,” a realistic training and simulation laboratory. There, researchers are investigating how experiential knowledge can be effectively, practically, and safely shared in the future.
 

ROBO.KIWI infographic showing how nursing expertise is captured with wearable sensors, analysed through AI, and transferred to humanoid robots for interactive nursing education and training.
© Fraunhofer IIS / KI-gestützte Visualisierung

Frequently asked questions about ROBO.KIWI

  • ROBO.KIWI is investigating how to preserve the experiential knowledge of nursing professionals over the long term and make it usable for training purposes. To this end, real-world nursing procedures are recorded and analyzed using mobile, body-worn sensors and then transferred to a humanoid robot system. The goal is to create a novel learning tool that not only documents nursing expertise but also allows trainees to physically experience it. In this way, valuable experiential knowledge will be preserved across generations.

  • Experiential nursing knowledge encompasses more than just technical and textbook knowledge. It also includes movement routines, interaction patterns, and situation-specific decisions developed over the years that nursing professionals apply in their daily work. An important component is what is known as embodied movement knowledge, such as during mobilization, transfers, or the ergonomic support of care recipients.

    This knowledge is often implicitly embedded in practical actions and is difficult to fully describe or document. When an experienced nursing professional retires, there is therefore a risk of losing knowledge and skills that are of great importance for high-quality care. 

    maphera Sensor Frontend
    © Fraunhofer IIS / Paul Pulkert

    To make this knowledge visible and analyzable, ROBO.KIWI uses the maphera® wearable sensor network. The mobile sensor platform captures movement patterns and physiological signals directly in day-to-day care and enables synchronous, multimodal data collection under real-world conditions. The data collected forms the basis for a dataset that can be used to analyze characteristic motion and interaction patterns and to make them available for nursing training.

  • Artificial intelligence analyzes the collected sensor data and identifies characteristic motion and interaction patterns of experienced nursing professionals. These patterns are consolidated into a digital knowledge model and transferred to a humanoid robot system. In this way, AI helps preserve experiential nursing knowledge and make it available for training purposes.

  • No. The humanoid robot in ROBO.KIWI is not intended to replace nursing staff, but rather to serve as an interactive learning partner in nursing education. It helps convey the motion sequences and experiential knowledge of experienced nursing professionals. Human skills such as empathy, communication, and individualized care will remain an indispensable part of professional nursing in the future.

    Instead, the robot is intended for use in so-called “Skills Labs" – realistic training and simulation labs for nursing care. There, trainees can practice practical skills under safe conditions and benefit from the experience of long-standing nursing professionals. The goal is to support nursing education and enable new forms of hands-on learning.

  • Nursing professionals acquire many important skills not only through textbooks or by observing others, but also through hands-on experience. While digital learning formats such as video applications, virtual reality (VR), and augmented reality (AR) have proven its effectiveness for teaching theoretical and procedural knowledge, they can only partially reflect the physical dimension of nursing interactions.

    This is exactly where ROBO.KIWI comes in: The humanoid robot is designed not only to demonstrate nursing expertise but also to make it physically accessible. In realistic training scenarios, the robot can demonstrate the motions of experienced nursing professionals or, as a care recipient, simulate realistic reactions and objections. This creates new opportunities to train practical skills under realistic and safe conditions and to establish proven patterns of behavior in the training process over the long term.

Find out more

 

Plug & Research

maphera®

Decentralized sensor network for mobile, multimodal physiological data collection

 

Motion Lab

Motion data for research

Objective analysis of gait and motion patterns for medicine and research

 

Strategic research projects

Research for the healthcare of tomorrow