Thomas Fleischmann developed an advanced method of measuring the cadence of road cyclists using neural networks as part of his final thesis at MCI with the support of his supervisor Bernhard Hollaus (Department Medical & Health Technologies).
The trend of tracking one's own sports activities is increasing and, road bikes considered a green alternative for locomotion in times of global warming. Therefore, cadence detection is becoming more vital and thus requires new measurement methods.
The classic measurement method existing is performed with a so-called "Hall Effect Sensor" on the bike frame and with a permanent magnet on the bike pedal. With each pedal rotation, the sensor on the bike frame is triggered. Over the time between two triggers, the cadence of the cyclist can be measured.
Thomas Fleischmann enhanced this classical method and was able to find a new way of measurement with the so-called inertial measurement unit combined with the fast-developing field of neural networks through his research in his master thesis. This new measurement does not have to be carried out with two sensors on the pedal and the crank of the road bike. It only requires a sensor at a previously defined point on the bike frame. The data measured by the enhanced method was then stored in a microSD and, after being processed, fed into a neural network. That made it possible to determine the cadence detection. In the study conducted in this research, four different road cyclists participated in obtaining data sets for neural network training and evaluation. So 10 hours of cycling data have been recorded to train the neural networks. The accuracy of this new method is approximately 90% in the training and the evaluation phase. Thomas Fleischmann was thus able to prove through his research that the cadence can be reliably measured through the inertial measurement unit and the subsequent feed into a neural network.
In an outlook on research, Thomas Fleischmann and his supervisor Bernhard Hollaus talk about additional possible applications of the cadence sensor. For example, the developed sensor can be used to measure and evaluate many other data of interest to athletes. As an example, cornering and inclines are mentioned.
For further questions about the research or the master thesis, please contact
Thomas Fleischmann (Research assistant am MCI) th.fleischmann@mci4me.at or Bernhard Hollaus (Lecturer in the Department Medical & Health Technologies) bernhard.hollaus@mci.edu
Thomas Fleischmann: Graduate at the Department of Mechatronics. He is currently working on his master thesis in the field of neuronal networks further he is employed as a Research Assistant at MCI. ©Thomas Fleischmann
