Neural networks for modelling and control of dynamic systems. A practitioner's handbook (Q1972829)

The main goal of the book is to describe approaches to neural networks (NN) based control. Then system identification is an integral part of such a control design. It is necessary to establish a model based on which the controller can be designed. For this reason the theoretical explorations in the book have been split into two main sections: a section about the system identification and a section about design of control systems as such. Ch. 1 introduces the multilayer perceptron NNs and discusses why and when it is relevant to use them in system identification and control system design. Ch. 2 outlines a procedure for system identification with NNs. Then Ch. 3 provides an overview of a wide range of approaches to NN-based control system design. Designs have been divided into two categories. In the first one the controller is itself a NN. In the indirect design, the designs are based on a NN model of the system to be controlled. The last Ch. 4 describes four case studies for the use of NNs for system identification and control. The book is written so to speak in a pragmatic way. This means the authors use a necessary minimum of description to get the required optimum for the controller design. The book is well readable and primarily addressed to engineering students at the graduate level. It would be, according to the authors, useful as a handbook in practical projects and courses. But this is not so clear as the book does not cover all subjects of applications of NNs in control.