MODEL CONVERSIONS OF UNCERTAIN LINEAR-SYSTEMS VIA THE INTERVAL PADE APPROXIMATION METHOD

This paper presents a new interval Fade approximation method to conver t a continuous-time (discrete-time) uncertain linear system to an equi valent discrete-time (continuous-time) uncertain model via interval ar ithmetic operations. Based on the inclusion theorem related to the int erval arithmetic, the interval Fade's approximants and their associate d interval error matrices with interval arguments are obtained via the Fade's approximants and their associated error matrices with degenera te (real) arguments, respectively. Tighter error bounds of various app roximate uncertain models with respect to their exact uncertain models are determined and used to modify the obtained Fade's approximants, s o that the resulting approximate uncertain models are able to tightly enclose the original uncertain systems. Thus, the analysis and design of the original uncertain systems can be indirectly carried out using the converted uncertain models in either the continuous-time or the di screte-time domain.