Author:

Daniel Homm
Supervisor:Prof. Gudrun Klinker
Advisor:Christian Eichhorn (@christian.eichhorn@uni-a.de)
Submission Date:[created]

Abstract

This study explores the potential of using wrist-worn inertial sensors to automate the labeling of ARAT (Action Research Arm Test) items. While the ARAT is commonly used to assess upper limb motor function, its limitations include subjectivity and time consumption of clinical staff. By using IMU sensors and MiniROCKET, a time series classification technique, this investigation aims to classify ARAT items based on sensor data. The dataset includes recordings of 45 participants performing various ARAT tasks. Results show that MiniROCKET offers a fast and reliable approach for classifying ARAT domains, although challenges remain in distinguishing between individual items. Future work may involve improving classification through more advanced machine learning models, better data augmentation. A different, but also imaginable approach would be the automatization of the ARAT with the help of augmented reality tools.

Results/Implementation/Project Description

Conclusion

[ PDF (optional) ] 

[ Slides Kickoff/Final (optional)]