Aplicação de sensores inerciais para detecção e classificação de manobras de skateboard

Abstract

The practice of the skateboard sport consists, in part, in the execution of skateboard tricks, which includes aerial tricks that are classified by jumps, adjustments and rotations of the skateboard in the air, which demand experience and visual acuity in order to be correctly classified. This document describes the development and discusses the experimental results of a project for application of inertial sensors (accelerometers and gyroscopes) on a skateboard. The capture and classification of these tricks based on the sensors’ readings are done by a computer program capable of receiving and analyzing the sensors’ data. The project was capable of correctly classifying 5 basic tricks performed on the skateboard by hand, as a proof of concept of the project’s functionality for classifying a real trick. As classifiers for the tricks, the project uses the overall shape of the inertial curves, the skateboard’s number of rotations in each axis in relation to the ground and the skateboard’s number of rotations in its transversal axis.