Adaptive fuzzy decentralized optimal control for interconnected nonlinear systems with unmodeled dynamics via mixed data and event driven method

In this work, based on the adaptive critic learning algorithms and fuzzy logic systems, the decentralized optimal control problem for interconnected nonlinear systems with unmodeled dynamics is investigated by applying a mixed data and event driven method. By introducing a modified performance index function for each auxiliary subsystem, the considered control problem of the whole interconnected system can be transformed into a group of event-driven optimal control problems of the auxiliary subsystems. Then, a three-layer neural network (NN) identifier is employed to reconstruct the completely unknown functions. Based on the obtained data-driven model, a single-critic NN is developed to solve the modified HJB equation. Meanwhile, a novel adaptive controller is designed by fuzzy logic systems to deal with the dynamical uncertainties cased by the unmodeled dynamics. Finally, a robot arm example is given to verify the effectiveness of the proposed control method.