Estudo do emprego de redes neurais artificiais para a classificação de padrões mastigatórios de caprinos

Abstract

This undergraduate final project describes the use of Artificial Neural Networks for classification of chew patterns of goats. It was considered the chew patterns of five different elements: no material in the mouth, two Plasticines of different textures, hay and oat. Information about patterns were obtained through the use of optical sensors based on fiber Bragg gratings sensors (FBG). This sensor monitors the biomechanical forces involved in chewing process. The interest of this approach is test the ability of neural networks to classify and recognize the presence of different elements during the chewing process of a goat. It was used a feedforward neural network based on the backpropagation algorithm for training.