Self-configuring, ambulatory motion capturing and analysis
Objectives and approaches
The aim of this line of research is to provide a system based on wearable sensors (inertial measurement units) that captures relevant kinematic, kinetic, and spatiotemporal movement parameters in real-time and measures these in an objective, reliable, and valid way. Based on this, the system should be able to provide context- and user-adapted analyses, while context here, for example, refers to the user’s current activity, i.e., the person wearing the sensors. At the same time, the system should be flexible, i.e., adjustable to different body parts, parameters, or individual reference values, and easy to use. In order to increase the usability and reduce errors, one focal point is the provision of a self-configuring system that automatically adjusts itself to the anthropometric characteristics of the user while also adjusting itself to the sensor positioning relative to the user’s body. To achieve this, new models, methods, and systems are being developed, implemented, and evaluated. These use model-based probabilistic estimation, optimal control, and machine learning approaches.
Design, implementation, and evaluation of new models and methods are carried out in cooperation between mathematics, computer science, movement science, and control engineering. The requirements analysis as well as the design and the evaluation of studies are supported by psychologists and cognitive scientists. Research questions concerning the extraction of anthropometric information from camera images as well as questions concerning machine learning are investigated in cooperation with the Department Augmented Vision at the German Research Center for Artifical Intelligence (DFKI). For questions regarding the connection of lightweight and reliable body sensor networks, we also cooperate with the Department Augmented Vision at DFKI as well as with the Microelectronic Systems Design Research Group at Technische Universität Kaiserslautern (TUK). For biomechanical issues, we cooperate with the Department of Movement and Training Science at TUK as well as with the Klinik Lindenplatz. Regarding medical and application-specific questions, we also cooperate with the Klinik Lindenplatz and with the University Hospital Halle (Saale). Research questions regarding the coupling of sensor measurements and optimal control methods are investigated in collaboration with the Chair of Applied Dynamics at the University of Erlangen-Nuremberg. Mathematical modeling issues are investigated in cooperation with the research group Technomathematics at TUK. In cooperation with the Chair of Control Systems at TUK we develop methods for motion prediction in the context of human-robot-interaction.