Physiotherapy Assistance

Physiotherapy is a form of specific training, which is primarily intended to restore, improve or maintain the mobility and functionality of the human body. While healthy people can automatically and unconsciously access all the building blocks of human erection and locomotion in their everyday orientation, the innate movement patterns for grasping and handling, turning and standing up, walking and running are only available to a limited extent when the central nervous system and the postural and locomotor apparatus are damaged. With the so-called reflex locomotion (or reflex locomotion), Prof. Vaclav Vojta has developed a method that makes elementary movement patterns accessible again, at least in partial areas, even in people with damaged central nervous system and locomotor system.

In cooperation with the Social Pediatric Center of the DRK Pediatric Clinic Siegen, we are developing, for example, in the project SenseVojta, a sensor-based system to support diagnostics, therapy and aftercare according to the Vojta principle described above. Our technical solution can support both therapists and parents in the implementation of therapy. For this purpose we design an electronic inertial measurement unit (IMU), which is equipped with several sensors. We use four IMUs that are placed along the arm to record the subtle stimulations. The integrated sensors record some medically relevant parameters of the patient as well as the movement parameters in 3D space. These data are then interpreted and documented with intelligent evaluation software regarding Vojta diagnosis and therapy. Especially for affected children, this solution brings a therapeutic added value, since on the one hand the Vojta principle is much more relevant for children than for adults and on the other hand only the use of a minimal-intrusive sensor technology is possible for children.

Our machine learning algorithms use the information collected by the sensors and help the therapists to identify the results of the therapeutic procedures. After preprocessing the data set, we generated features according to the so-called codebook approach. Inspired by word processing, we further developed the codebook approach towards modeling and classification of multimodal time series. The training and test data are divided accordingly in a hold-out procedure. We use the Support Vector Machines to classify the different activation levels. Currently, our described pattern recognition platform achieves a classification rate of about 85% in the recognition of four stimulation levels.

Contact person

Prof. Dr.-Ing. Marcin Grzegorzek

Third party funded projects and publications

BMBF-Project: SenseVojta - Sensor-based Diagnosis, Therapy and Aftercare According to the Vojta Principle. Duration: 01.12.2016 - 30.11.2019.

Khan, M.H.; Schneider, M.; Farid, M.S.; Grzegorzek, M. Detection of Infantile Movement Disorders in Video Data Using Deformable Part-Based Model. Sensors 2018, 18, 3202.