Fuzzy controllers design based on neural beam dynamics model optimized by genetic algorithms

This work proposes a methodology of fuzzy controllers design. They are obtained by an optimization process that uses genetic algorithms. For this optimization procedure, the knowledge of the system dynamics is required. Therefore an artificial neural network is trained to model the dynamic behavior of the plant from the experimental inputs and outputs of the system. The rule base, the weights of the rules and the input membership functions are optimized. The proposed methodology is evaluated experimentally on a steel cantilever beam controlled by piezoelectric actuators. Those controllers are evaluated on time and frequency domain. The obtained results confirm the efficiency of the proposed methodology.