Smartwatches for Sport Applications


In this project we - a group of 13 students (electrical engineering) - investigate the usage of so-called smartwatches for sport applications. Our general goal is to get accurate motion information out of the data we can collect through smartwatches. Most smartwatches these days have all kinds of sensors, but especially interesting for our motion analysis are the gyroscopes and accelerometers. With this data we hope to capture how you move and then for example recall what you did or how well you’ve been doing. Every sport needs a different approach, so we’ve split up in teams to optimise what we know best and bring all of our knowledge into the analysis.

Participating Students

  • Alexander Weber
  • Dawyd Klimaschewski
  • Ruben Fiedler
  • Julian Jebe
  • Johannes Hoffmann
  • Daniel Beus
  • Fabian Bauer
  • Jannek Winter
  • Tim Rocholl
  • Michelle Djomo Njamen
  • Solveig Baschin
  • Kristina Apelt
  • Michel Boldt


  • Marco Gimm
  • Bastian Kaulen
  • Gerhard Schmidt




In this part of the project we examined swimming sport by using the sensors of the smartwatch. In the framework of this project the goal was to find a method to discriminate between the four official swimming styles defined by FINA (abbreviates Fédération Internationale de Natation):

  • breaststroke,
  • front crawl,
  • backstroke, and
  • butterfly.

At first we collected data to create a database for the four types of swimming styles. Therefore, we recorded data of different swimmers as well as different distances using the smartwatch to get a detailed background of information. By analysing the video of the vector movement of the acceleration we noticed some characteristic pattern. Based on these pattern we chose the acceleration for the comparison.

To have more information we additionally used the information about the rotation because it also differed much in the style and has repetitive values compared to the remaining sensor records. To use the records in the following steps the files were cut to get the relevant data without the running idle at start and end. Afterwards the variances of the three dimensional vectors of the acceleration and rotation were calculated to have the possibility of comparing them with other records. 60 percent of the record sets were used to serve as a reference in the program by determine the averages of the variances.

Using MATLAB we created a program according to the principle of codebook which uses the vector consisting of the averaged variances of rotation and acceleration as a reference for every swimming style. With these references we tried to analyse the swimming stroke of another recording by comparing its values of acceleration and rotation. By calculating the norm of each reference with the vector of data which should be examined we got comparative values. The program selected that swimming style, which norm had the smallest value and, thus, the highest accordance.

Testing the program with the remaining 40% of the records showed that it selects always the right swimming style, except for backstroke. Because of the similarity in the variances to front crawl the algorithm is not capable to determine the right decision, which was expected due to the alikeness of the two swimming styles. To solve this issue, other features had to be considered in future research.

Team members
  • Solveig Baschin
  • Kristina Apelt
  • Michelle Djomo Njamen
  • Michel Boldt
Example Measurements


Animation of the specific movement pattern of the swimming styles over time.


Prof. Dr.-Ing. Gerhard Schmidt


Christian-Albrechts-Universität zu Kiel
Faculty of Engineering
Institute for Electrical Engineering and Information Engineering
Digital Signal Processing and System Theory

Kaiserstr. 2
24143 Kiel, Germany

Recent News

New PhDs in the DSS Team

Since January this year we have two new PhD students in the team: Elke Warmerdam and Finn Spitz.

Elke is from Amsterdam and she works in the neurology department in the university hospital in the group of Prof. Maetzler. Her research topic is movement analysis of patients with neurologic disorders. Elke cooperates with us in signal processing related aspects of her research. Elke plays ...

Read more ...