Proceedings Sportinformatik 2012 1 9. Symposium der dvs-Sektion Sportinformatik, Univ. Konstanz, 12.-14.9.2012 Self-Aligning and Drift-Compensated Rowing Seat Position Measurement System Based on Accelerometers and Magne- tometers Franz Gravenhorst¹, Christoph Thiem², Bernd Tessendorf¹, Rolf Adelsberger¹, Christina Strohrmann¹, Bert Arnrich¹, and Gerhard Tröster¹ 1 Wearable Computing Lab, ETH Zurich, Switzerland, {lastname}@ife.ee.ethz.ch 2 Institute of Numerical Methods in Mechanical Engineering and Graduate School of Computational Engineer- ing, Technische Universität Darmstadt, thiem@fnb.tu-darmstadt.de Introduction and Related Work The rowing motion is a complex sequence of limb movements. In order to improve the boat speed as well as to prevent injuries, athletes and coaches are interested in analyzing and optimizing it in an objective and reliable way. In previous work, we implemented a sensor network to measure the oar movement and the boat stability [11, 4, 5]. In this work we focus on extending our system with a position tracker of the sliding seat. The movement of the sliding seat is directly linked to the rower’s leg movement which involves the rowers’ most powerful muscles and thus is a key element in row- ing technique. Bad technique in the leg movement significantly slows down the boat speed and can result in injuries. The most commonly used system to monitor the sliding seat movement nowadays is the commercial StrokeCoach TM device by Niel- sen Kellerman [1]. This system consists of a permanent magnet attached to the sliding seat and a reed switch fixed to the boat. This way, it measures the stroke rate (how often the rower moves the seat back and forth per minute) but there is no information about the shape of the motion. Another approach is presented by Kleshnev [8] and Smith et al. [9]. They attach a cord to the sliding seat which drives a potentiometer when the seat moves. These measurements are proven to be help- ful for coaches. However, the installation and calibration requires some effort and elite rowers report unnatural feelings as the cord constantly pulls the seat to one di- rection. An approach based on optical methods is presented by Davoodi et al. [3]. They employed the system and obtained reasonable results for indoor rowing. However, it is not yet ready for on-water environments. Contribution We present a seat position tracker system which consists of two miniaturized measurement units attached to the sliding seat and the boat shell. Each unit fea- tures a tri-axial accelerometer and magnetometer. We present a method how to process the raw data to calculate the seat position. The magnetometer data is used to correct for low-frequency sensor drift. Three accelerometer axes and prior knowledge concerning the boat geometry is exploited to feature a dynamic calibra-