An Algorithm for Kharitonov Synthesis

A multiple plant model formulation with a quadratic cost function has been shown to be a method for synthesizing stabilizing feedback controllers based on Kharitonov's theorem. An algorithm which solves the multiple plant model quadratic cost function problem is presented in this paper. The algorithm starts with a different full state feedback gain for each model then attempts to converge each gain to the fixed Kharitonov feedback gain. The Kharitonov feedback gain can have an output feedback or decentralized output feedback structure. An extension of the algorithm to synthesizing optimal gain scheduled control laws over a regime of plant models for centralized and decentralized output feedback is also presented.