An augmented system based on machine learning for boccia assisted gameplay

Abstract

In order to promote the practice of sports, several approaches using technology have been employed to gamify and augment the user experience. Following this trend, the research group proposed an approach to encourage the practice of Boccia, while promoting social inclusion and reducing the amount of time it takes for newcomers to the sport to become proficient by gaining knowledge of game tactics. The present work focus on the detection, in real-time, of Boccia gestures for the framework proposed in a previous work by using a wearable device to detect the gestures. To evaluate the correct functioning of the system, several types of tests were carried out. First, the developed machine learning model was evaluated in terms of accuracy, recall, among others. Then, the gesture detection system was tested with 15 participants that executed the different Boccia gestures while using the wearable placed on the wrist. Finally, tests were carried out to integrate the gesture detection module into the framework proposed in a previous work. The tests yielded positive results that allowed the validation of the use of the system in the Boccia game.