Uma abordagem de fusão de dados multisensores em dispositivos IoT para jogos sérios

Abstract

This work presents a sensor fusion approach in the realm of serious games and internet of things. For this I tried to use important sensors that are present in a smartphone (accelerometer and gyroscope). For this, an app was created that captures the data of these sensors on the smartphone. The methodology is divided into: input, processing and result. At the entrance, the user simulates the movement of a tennis move, the data from the gyroscope and accelerometer are stored in a file with a predefined layout, these data are used in the processing step. During processing, the file with the sensor data is processed, removing unnecessary information and the sensor information is merged by a complementary filter, and the information from two sensors becomes one. In the results, the user’s movement data is submitted to a KNN classifier that already has a base with professional movement movements, and a classification is made for each point generated. Then, an estimate of the proximity of the movement of the amateur user to the professional is presented.