Controle tolerante de quadricópteros em cenários com falhas em atuadores e sensores

Abstract

The methods of Fault Detection and Diagnosis (FDD) and Fault Tolerant Control (FTC) have attracted significant research attention due to the impositions of technological development of different systems to ensure safe operations, continuity of execution of required functions, cost reduction and optimization of maintenance tasks. Unmanned aerial vehicles (UAVs) are examples of these systems, especially the quadcopter. Due to its characteristics, the quadcopter is currently applied in large scale to solve several problems and adversities. In view of this, this work implements different FDD approaches based on Kalman filter (linear, extended, and unscented, all with a three-stage and adaptive structure) for fault estimation, and a tolerant control for a quadcopter, considering different faults with nonlinear behavior and with the possibility of simultaneous occurrence in its actuators and sensors. The disturbances caused by winds are addressed in the analysis of the FTC system behavior, together with others obtained by comparing the performance of different versions of the Kalman filter. Among the results, the FDD methods were efficient in estimating the faults in the considered scenarios, especially the adaptive filter. This provided a significant improvement in the safety and reliability of the quadcopter by the FTC structure, as the system that previously became unstable in the presence of the faults, remained stable with the implementation of this type of control.