Self-learning fuzzy control strategy of two-layer networked learning control systems based on improved RBF neural network

This paper is concerned with two-layer networked learning control system architecture that is consisted of local controller and learning agent. Firstly, networked nondeterministics are tracked respectively by zero order holding (ZOH) and cubic spline interpolator in local controller and learning agent. Then fuzzy control strategy is used in local controller, and an improved radial basis function (RBF) neural network by combing the regularized parameters with the leave-one-out cross-validation criterion is employed in learning agent. Taking advantage of fuzzy control and RBF neural network, a self-learning fuzzy control method is proposed to improve the control performance, where RBF neural network is used to dynamically tune the parameters of local fuzzy controller. Finally, simulation results confirm the effectiveness of the proposed scheme.