Generic signature generation tool for diagnosis and parametric estimation of multi-variable dynamical nonlinear systems

In this paper, a graphical signature generation tool is proposed for diagnosis and parametric estimation for nonlinear systems. This tool is based on the definition of a two-dimensional graphical signature obtained by using the history of the output measurements on some moving-horizon window. The underlying idea is that if each parametric variation deforms this signature in a different way, the latter becomes a remarkable tool for diagnosis, in particular in the case of the diagnosis by human operator. The whole scheme is illustrated by a simple example.