Least squares approach to fictitious reference based tuning of full parameterized linear time-invariant controllers

In this paper, a parameter tuning method of full-parameterized linear time-invariant controllers is presented. Our approach is based on the direct use of the experimental data without mathematical models. Concerned with this point, the author and his colleagues have proposed a so-called “fictitious reference iterative tuning” method (FRIT) where full-parameterized controller is tuned in the off-line nonlinear optimization with only one-shot experimental data. This paper expands the previous method to be performed in the off-line least squares.