Semi-Empirical Continuous Time Neural Network Based Models for Controllable Dynamical Systems

Abstract

We discuss the problem of mathematical and computer modeling of nonlinear controllable dynamical systems with incomplete knowledge about the object of modeling and the conditions of its operation. The suggested approach is based on a merging of theoretical knowledge for the system with training tools of artificial neural network (ANN) field. We present an extension of previously proposed semi-empirical neural network modeling methods for the case of continuous time ANN-models, which makes it possible to expand the possibilities of this approach. The efficiency of this approach is demonstrated using the example of motion modeling for a maneuverable aircraft.