The IFAF funded project "KI auf biomech. Simulation zur Bewegungsanalyse in der Therapie" (BIOKIT) is a joint project of:

• Berlin University of Applied Sciences working group softrobotics of Prof. Dr.-Ing. Ivo W. Boblan,
• Hochschule für Technik und Wirtschaft Berlin working group medical physics of Prof. Dr. Stephan Matzka,
• Fraunhofer Institut für Zuverlässigkeit und Mikrointegration e.V..
• trinitas Therapie.Bewegung.Sport. GmbH.

Musculoskeletal disorders are the most common cause of incapacity to work. At the same time, there is a supply gap in the healthcare industry. Digital applications based on motion analysis address this problem. The potential for technical development in these applications lies in the reliability of the feedback generated in real time. Taking this up, the project is developing a demonstrator that analyzes exercise execution for a therapeutic movement program and gives users real-time feedback.

To achieve these goals, the following components work together:

• multimodal tracking system,
• biomechanical simulation of musculoskeletal models,
• automated process chain and
• machine model trained for the application.

Supervised learning methods are used to develop machine models based on the tracking data and biomechanical simulation that estimate the results of motion analysis in real time. The reliability of the movement analysis and feedback is examined using the automatic process chain. This determines the confidence of the demonstrator for use in therapeutic practice: It is assessed whether the movement analysis is sufficiently reliable to generate correct feedback. Furthermore, approaches for optimizing the movement analysis and the tracking system are derived.