Cost-efficient system for movement quantification in neurology
Research Area:Clinical Movement Control and real-life Behavior Analysis for Assistive Systems
Researchers:Nicolas Ludolph; Björn Müller; Martin A. Giese; Winfried Ilg
The quantification of neurological deficits, especially in gait, is important for many applications in diagnosis as well as rehabilitation training. Standard movement analysis systems (VICON, MotionAnalysis, ..) are extremely expensive and not suitable for home deployment. The development of the Microsoft KINECT sensor has made it possible to realize body tracking with a cost of a few hundred Euros. We exploit this technology for the development of cheap movement analysis systems that cost less than 3000 Euros and which can be deployed at multiple sites and in principle also at home environments. The developed system fuses the information from several KINECT systems in order to allow the analysis of large-range motion, such as walking. The system is presently exploited in studies on movement deficits in cerebellar Ataxia, and was exploited already for a field study in Holguin in Cuba. Further applications are multi-center studies, permitting a distribution of the same hardware for movement analysis at a low price to multiple centers for the study of rare neurological diseases.
The developed system consists of six Kinect™ v2 sensors with attached mini- PCs, host computer. The mini-PCs are time synchronized with the host PC and are communicating via WLAN. Using this architecture we can scale up to as many sensors as necessary to cover a given volume. For the gait analysis, we used six Kinect sensors covering a walking corridor of 6-7 meters. Performance is evaluated by comparison with data from a high-end motion capture system (VICON).