Chapter 3 - Neural Networks: A Methodology for Modeling and Control Design of Dynamical Systems

Abstract

This chapter presents different approaches where neural networks (NNs) have been successfully employed for the modeling and control of nonlinear dynamical systems. NNs have allowed to deal with complex and nonlinear systems, which are assumed to be uncertain, disturbed, unknown, or difficult to model; therefore, the resulting neural models become robust schemes against such uncertainties. The flexibilities of the NN are exploited and highlighted in the designs with the aim of proposing well-established models, which guarantee the system properties and with mathematical structures such that convenient control methodologies can be synthesized. This chapter covers both the discrete-time and the continuous-time framework for modeling and control of uncertain systems, where recurrent NNs (RNN) are used to develop the artificial neural models, which a posteriori are employed for the design of two neural control schemes (sliding mode block control and nonlinear optimal control). Finally, the applicability of the neural methodology is illustrated through two practical examples.