Periodic Extended Kalman Filter to Estimate Rowing Motion Indoors Using a Wearable Ultra-Wideband Ranging Positioning System

Rowing depends on weather conditions and athletes rely heavily on dryland training, specifically, on rowing machines. Past research implemented methods using inertial measuring systems to study the rower’s biomechanics and oar dynamics, mostly for outdoor rowing. Tracking the oar dynamics is of great significance to coaches and athletes because it is correlated to the rower’s technical skills and boat speed (on-water). Ultra- wideband (UWB) ranging as an indoor localization method is one of the most accurate and precise among available radio frequency technologies. UWB has been used in game and individual sports to track athletes and their performance during training and competition. This paper proposes an UWB indoor positioning system to track the periodic motion of a rowing machine’s handle at varying frequencies. Furthermore, it introduces a periodic extended Kalman filter to estimate the motion as a non-sinusoidal wave. The proposed UWB positioning system is based on the double sided two-way ranging time of arrival between two anchors and a single tag. The ranging measurements are processed using three mathematical models, trilateration, extended Kalman filter with constant velocity, and a periodic extended Kalman filter. The results are compared to a reference trajectory obtained from a Vicon 8 camera motion capture system. The periodic Kalman filter outperforms the other two models, however, the accuracy and precision of the predicted position of the handle depends on the accuracy of the ranging measurements. The proposed indoor rowing positioning system is able to effectively track rowing motion using UWB ranging and a periodic extended Kalman filter.