Parametric Identification of Nonlinear Systems by Aggregation of Static and Dynamic Neural Networks

The parametric identification of nonlinear systems opens up broad prospects for the creation of modern monitoring and control systems, thereby ensuring a significant increase in their technical and economic indicators. The paper uses methods of mathematical modeling, numerical methods and the theory of neural networks. The result of the work was the proof of the possibility and acceptability of using a combination of dynamic and static neural networks for the tasks of identifying nonlinear systems. It is preferable to use a multidimensional and multilayered direct transmission network as a dynamic neural network, and a radial basis network as a static network.