Estudo e desenvolvimento de sistemas de controle com filtragem estocástica aplicados em um drone quadricóptero

Abstract

The control theory and signal filtering, especially regarding their relation to nonlinear system control problems, have always been considered by many mathematicians and engineers as theories of broad application in combination. This present work provides motivation for the study of these areas, approaching the concepts of control theory and stochastic filtering in a didactic way. For this approach, a prototype of a quadcopter UAV (Unmanned Aerial Vehicle) was assembled só that one of its control axes serves as a system where control and filtering techniques will be tested and validated. To this end, a platform is proposed to fix the UAV só that the control axis can be identified in a controlled and safe way. Black box and white box system identifications were used for the development of the controller and Kalman filter, respectively. Frequency analysis and indicators were used to compare different control techniques applied in the work. With the application of the Kalman filter on the Arduino, it was possible to assemble a feedback control loop with reliable feedback. Both control techniques compared here were able to stabilize the naturally unstable system.