CONTROLE PID NEURO-ADAPTATIVO EMBARCADO EM UM SISTEMA MAGLEV
CONTROLE PID NEURO-ADAPTATIVO EMBARCADO EM UM SISTEMA MAGLEV
-
DOI: https://doi.org/10.22533/at.ed.9282428044
-
Palavras-chave: Microcontroladores, Controle adaptativo, Redes Neurais Artificiais, MAGLEV.
-
Keywords: Microcontrollers, Adaptative Control, Artificial Neural Network, MAGLEV.
-
Abstract: ABSTRACT: The present paper proposes a PID controller tunning method based on an artificial neural network, implemented in an embedded microcontroller system. The neuro-adaptative controller architecture is known for its efficiency in non-linear models, with constant parameter changes, in this case, a MAGLEV device. Use largely spread in industrial fields such as: high-speed magnetic levitation trains, magnetic bearings, vibration dampers etc. The research will be presented in the following sequence: study of the non-linear model, definition of the neural network architecture, evaluation of the hardware specifications, and with that data, the algorithm of the neuro-adaptative controller will be developed, alongside from the firmware. Completed those steps, the test and implementation stage of the proposed control method will begin, reaching its validation through comparison with a classical PID controller. All resources built from this project will be available to all the community in GitHub.
- Eduardo Marek
- Alexandre Dalla'Rosa