On output stabilizations problem for linear continuous systems by a neurocontroller

In this paper, we study the capabilities of artificial neural networks in solving the problem of stabilizing the equilibrium position of a linear second-order differential system (in particular, of a mechanical system) by control depending on the retarded values of the output vector. A “half” of the state coordinates of the system (excluding the first derivatives) are used as the output vector. Using the model of an inverted pendulum on a cart as an example, the constructed control is compared with other types of controls.