On QFT tuning of multivariable mu controllers

Optimal control involves feedback problems with explicit plant data and per formance criteria for which a solution is either synthesized or ruled out. H-infinity optimal control is probably the most renowned technique in this class where the control synthesis procedure involves various iterations ove r weightings. In this paper we argue that the integration of optimal contro l synthesis and manual tuning in the quantitative feedback theory (QFT) des ign environment enables design of controllers with levels of performance th at surpasses what can be achieved using only a single technique. Specifical ly, using a constructive example, we demonstrate that QFT's open-loop tunin g can be more transparent than tuning closed-loop weights. In this example, QFT tunes the mu controller with the objective of reducing control bandwid th while maintaining robust performance (mu < 1). (C) 2000 Elsevier Science Ltd. All rights reserved.