Fraunhofer Research Institution for Individualized and Cell-Based Medical Engineering
Fraunhofer Research Institution for Individualized and Cell-Based Medical Engineering
Foto: © Olaf Malzahn

AI Motion Labs - Motion Analysis

Movement Reimagined – Using AI for Better Rehabilitation

In the AI-Motion Labs, movement is captured, analyzed, and therapy-oriented using the latest sensors and artificial intelligence. The goal is individualized, technology-assisted rehabilitation with transferable relevance to everyday life.

The AI-Motion Labs at Fraunhofer IMTE represent a highly specialized research environment for capturing, analyzing, and interpreting human movements – with a focus on medical rehabilitation and therapy-assisted technologies. In close cooperation with the University of Lübeck, two labs have been set up, equipped with a perturbation treadmill and a 3D rope robot for body weight unloading.

The systems interact in a virtual reality environment and enable automated movement analysis using specialized camera technology. The sensor data captured are converted into digital body models, which are evaluated using artificial intelligence techniques to identify patterns, deviations, and therapy outcomes.

A central goal is to transfer robotic-assisted assistance systems into the rehabilitation routine. While many data-driven systems already exist in clinical research, robust approaches to interpreting the data collected are often lacking. This is where the AI-Motion Labs come into play: by combining analytical biomechanical models and data-driven AI systems, therapy efficiency can be enhanced and individualized. 

Fields of research

Expand all Close all

Data-Driven Assistance Systems in Rehabilitation Robotics

Most robotic-assisted assistance systems capture a wide range of data, necessary for the functionality and safety of the devices. However, these data often only allow conclusions about physiological conditions under very specific assumptions and simplifications. Yet, certain patterns in the data can often be identified, which cannot always be described using analytical or numerical methods. Using data-driven approaches, these correlations can be captured and evaluated. The research focus is on describing the relationships between device measurements and physiological parameters. Furthermore, prediction of rehabilitation success from metadata or assessment data is aimed at.

Model-Driven Assistance Systems in Rehabilitation Robotics

The assumption that human movement can be reduced to the interaction between segments, joints, and actuators allows for a model-driven approach. The goal of this research is to map patient-specific movement disorders into a generic movement model. Through subsequent forward simulation, potential interventions can be assessed, and therapy-specific milestones determined. This analytical approach provides deep insights into the relationships between movement disorders and physiological parameters.

 

Knowledge Transfer into Rehabilitation Practice

Precise and individualized movement analyses and therapies, as well as automated training evaluations, offer the opportunity to save resources in rehabilitation practice and relieve medical staff. For example, robotic assistance systems could support the redevelopment and implementation of standardized assessments or make patient-specific corrections during rehabilitation to ensure optimal recovery.

Services

  • Development of Data-Driven Models to Link Device Information with Movement Parameters
  • Development of patient-specific movement models
  • Planning and conducting clinical trials in the AI Motion Labs
  • Planning and conducting studies for product validation of assistance systems in rehabilitation