Development of Signal Processing and Machine Learning Methods for Inertial Sensor Based Motion Feedback Systems
Event date:  July 7, 2016 3:30 PM  to 4:30 PM

Talk Abstract:

The provision of motion feedback is known to be a decisive variable for motor skill acquisition and motor learning. However, despite the development of wearable motion sensing technologies, augmented motor support did not become common practice in sports training so far. This work develops signal processing and machine learning methods for the provision of ubiquitous motion feedback on the base of inertial measurement devices. Precisely, I will present a system for the assessment and rating of motion style. This mobile, computer-directed motion feedback application is subject to four main procedural stages. First, numeric motion data for subsequent machine data processing are collected from the inertial sensors. Second, the information content of the acquired motion data is augmented to provide accurate and reliable kinematic motion information. Third, the augmented motion data is transformed so that meaningful data representations are created. Lastly, artificial motion knowledge is learned and then utilized to enable the retrieval of relevant motion information for feedback provision to the user. Every computational stage is illustrated with practical motion data from ski jumping, and could serve for other sports in a similar way. This work constitutes an important contribution to the implementation of future training and motion feedback software tools, supporting multiple aspects of a motion performance. Especially for judging-based sports, the presented intelligent style assessment could provide fundamental and unique information to increase objectivity and measurability of the final competition scores.

 

Speaker’s Bio

Heike Brock is a PhD candidate at the Graduate School of Media and Governance, Keio University, Japan. Her research interests focus on the development of motion information retrieval technologies for motion feedback tools from wearable sensor data. Before coming to Japan, she has been a research member at the Institute of Sport Science at Leibniz University, Hanover, where she implemented a wearable movement sonification system for use in motor learning and rehabilitation.

Heike received a M.Sc. in Visual Computing from Saarland University in 2011 and a B.Eng in Audiovisual Media from Stuttgart Media University in 2008. During her master studies in Saarbruecken, she has been a student assistant in the Multimedia Information Retrieval group in the Cluster of Excellence for Multimodal Computing and Interaction (MMCI) at the Max-Planck Institute of Informatics, Department 4 Computer Graphics. She is currently a scholar of the DAAD and the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT).

  

  


Event location:  Universität Stuttgart
SimTech Building
Room 00.009
Pfaffenwaldring 5a
70569  Stuttgart