Sistema de processamento de imagens para o auxílio na obtenção de funções de transferências

Abstract

Several daily activities utilize control systems and they can be modeled and represented by the interconnection of several subsystems, providing a better understanding of the composition of their components. These, when used in conjunction with Transfer Functions (FT), also describe the cause and effect relationships of the whole system. Signal flow diagrams are directed graphs used to represent the causal relationship between system components, where their nodes and edges represent signals and FT, respectively. To calculate the final gain of a diagram, a widely used method is the Mason gain formula. This method requires the user to find all the direct paths and loops in the graph. All this processing has a relevant complexity, since it involves visual analysis and mathematical manipulations, making this method susceptible to errors. Thus, the aim of this work is to describe the development and implementation of an image processing system, capable of processing the image of a signal flow graph drawn on a paper or whiteboard and performing all necessary steps, until find all the paths present in it and the equivalent FT. In order to validate this approach and evaluate its accuracy, we obtained 117 samples of signal flow graphs with university students of the Electrical Engineering course at UFRN. The main contributions proposed in this work are to help both students and professionals in calculating the final FT with a higher probability of success, as well as to streamline this process.