Online Multivariable Identification and Adaptive Self-tuning PID Control of a Flying Boat System Using Neural Networks

In this paper, a new approach is proposed for online identification of the dynamics of a flying boat system based on neural networks (NNs). Moreover, a new NN-based self-tuning PID control scheme is developed for the purpose of the adaptive control of nonlinear systems. The neural network employed for this purpose is one of the novel ones called Growing and Pruning Radial Basis Function (GAP-RBF). The Extended Kalman Filter (EKF) learning algorithm is utilized for parameter updating in the GAP-RBF neural network. The proposed identification approach is applied to a multi-input, multi-output flying boat system, and its performance is evaluated. The obtained results show the excellent performance of the proposed scheme in exact modeling of the system.

کلیدواژه ها:

GAP-RBF ، EKF ، PID ، flying boat