Reconfiguração de controle em um drone quadrimotor auxiliada por um classificador de falhas

Abstract

The tasks of fault classification and reconfiguration of control in processes or systems are areas in whichresearch is in increasing development, and this expansion can be associated with its preventive charac-teristic that together with a robust control prevents both material and human losses. In this context, thiswork presents a new proposal of adaptive approach for classification of failures and active reconfigura-tion of control in dynamic systems, based on two steps: In the first step, the proposed approach uses classic methods for classifying failures; in the second stage, it proposes variations in control reconfigu-ration in dynamic systems from data of fault identification. To verify the efficiency of the approach, itwas applied in a four-engine drone model with failure in one of the engines, contextualized as the poweroutage in this. Thus, the results obtained, through simulation without Matlab software, by rewriting sys-tem are compared with controllers used only in normal operation, with an emphasis on varying height.In general, the proposed approach has obtained good results, suggesting that it is promising to performthe tasks of classification of failures and reconfiguration of control of dynamic systems.