Experimental design for the construction of regression models of controlled systems (Q1108718)

This is a review paper and surveys studies made on optimal experimental designs (ED) for regression models. The author has examined the traditional experimental design theory and best estimation of parameters when the noise is a \(\delta\)-correlated stochastic process. Details of ED and passive identification methods are also given. Studies on optimal ED in dynamical controlled systems and effective choice of signal testing have also been surveyed. Frequency-domain designs and time-domain designs are covered. Computer-aided control systems with applications to known ED methods for identification of dynamical processes are cited.