The utility of wearable sensors to identify movements and quantify training load in fencing

Abstract

The most common attacking movement in fencing is the lunge and its objective is to land a hit on the opposing fencer with the weapon while defending oneself in order to accumulate points. The fencer (or team) with the highest points at the end of the bout wins. Other commonly executed movements in fencing includes the step lunge and the fleche. In competitive sports, various research has been conducted to improve on athletes’ techniques and training methods with one such method being wearable sensor technology. However, research in wearable sensor technology and its applications in fencing is lacking. The objectives of this research project are firstly, to develop a system that uses wearable sensors to identify fencing-specific movements in real time namely: the forward step, backward step,lunge, step lunge and fleche. Secondly, we aim to quantify the intensity of each lunge and by extension,estimation of training load at the end of each training session or bout. The data will allow coaches and athletes to not only study the quantity and quality of fencing-specific movements performed during a bout but to also monitor the cumulative training load undertaken by a fencer. This objective data will allow coaches to better plan a periodized program for their athletes with the aim of enhancing performance and reduce the risk of injuries. This project uses a customized relative orientation sensor consisting of a three-axis accelerometer and a three-axis gyroscope which provides acceleration and angular velocity. The sensor is lightweight and small in size which allows for data collection without restricting the fencer’s movement. Data was continuously collected via the Bluetooth module at a range of 16m. A graphical user interface was developed to provide a training summary which displays the intensity and frequency of techniques specifically the forward step, backward step, lunge, step lunge and fleche which have been executed. The algorithm has been developed to analyse the different signals collected as each technique is being executed and categorise them as the forward step, backward step, lunge, step lunge or fleche. To provide feedback on the training workload, the acceleration and velocity peaks were used to compare with maximum peaks collected. With this interface, the fencer can improve their training efficiency.