Simulação de um sensor de força multiaxial utilizando métodos de análise digital de imagens e classificador neural

Abstract

This work develops a set of simulations about the operation of a digital multiaxial force sensor that performs the measurement of the forces applied to its structure, through techniques of digital image processing and analysis and classification by a neural classifier (Artificial Neural Network). The proposed sensor has a flexible structure that connects two concentric and parallelly arranged surfaces where one of the surfaces has a predetermined image and the other has an image sensor. The entire simulation is performed virtually via MatLab software, including the implementation of the digital analysis of the set of image samples and the classification of the direction and direction of the applied force. At the end, the results are analyzed and improvements are proposed for future projects.