Autoregressive Model for Identification of Nonlinear Dynamics Systems With the GNU Octave Control Toolbox

In this paper is shown how to implement the GNU Octave Control toolbox for autoregressive models for the identification of nonlinear dynamics system with the GNU Octave control toolbox. This paper is focused more on the linear autoregressive model which is based in the ARX structure (autoregressive external input model) for nonlinear dynamic system identification. This research study begins with the theoretical framework in which it is included the fundamentals of nonlinear dynamic system identification along with the definition of the z transform and the definition of the ARX structure for nonlinear dynamic system identification. The third section consists in the definition of the GNU Octave control toolbox that are implemented for auto-regressive model structure for nonlinear dynamic system identification. Then three numerical experiments are performed which consists in the following; first the identification by auto-regressive external model structure of a random signal and then the identification of a quadratic function. This numerical experiment section finalizes with the identification of the attitude nonlinear dynamics of a quadrotor unmanned aerial vehicle UAV, taking into account that the nonlinear dynamics of this UAV is complex because of it is highly coupled. With these numerical examples, it is evinced how to implement the GNU Octave Control toolbox function in order to perform an accurate dynamic model identification that provides a compact model in order to be used for adaptive control or only to be implemented to identify a data series with a reliable mathematical model. In this paper is show how to efficiently implement the GNU Octave Control Toolbox function describing the advantages in simplicity and efficiency for identification purposes. This paper concludes with the appropriate discussion and conclusion sections.