Quantitative design for SISO non-minimum-phase unstable plants by the singular-G method

Abstract

It has been shown how a single-input-single-output (SISO) plant can be converted by several independent samplings of the output, into an equivalent multiple-input-multiple-output one, with respect to feedback system stability. The singular-G method may then be used to stabilize the system despite plant uncertainty. Only qualitative properties were previously presented. This paper presents a quantitative design procedure for such uncertain SISO plants with right half-plane poles and zeros, and some of the costs thereby incurred. The inherently large loop sensitivity is transferred from the plant to the compensators. Plant disturbances are greatly amplified rather than attenuated. But for a large class of such uncertain plants, stabilization and reasonable command response are possible.