Nonlinear identification and control. A neural network approach (Q5949793): Difference between revisions

The monograph presents the neural network (NN) approach to nonlinear identification and control. Basically, the monograph consists of three parts. The first part (25 pages) gives a short overview to fundamental principles of NNs. Namely, various types of NNs, are described in a not very concise way. Learning algorithms and the problem of approximation are also mentioned. Then some applications possibilities of NNs are selectively enumerated. The main part of the monograph is devoted to methods for nonlinear identification (NLI) using NNs (100 pages). Namely, sequential NLI with variable NNs, recursive NLI with Volterra polynomial basis function NNs, multiobjective NLI, when model selection is made by genetic algorithms and finally wavelet based NLI are described in detail. In the third part (67 pages) various techniques for nonlinear control (NLC) using NNs are studied. Concretely, there are nonlinear adaptive and predictive controls.