Reduced order suboptimal linear quadratic controller using Krylov subspace projection

Abstract

The paper presents a method of designing a suboptimal linear quadratic controller for the reduced order approximant of a higher order system. The optimal control design for higher order systems is complicated because of the reason that for optimal control feedback from all the state variables is required. In this work, Krylov subspace projection is used to obtain the suboptimal controller parameters, based on the use of only measurable states. Krylov subspace projection is one of the most efficient solutions for reducing the higher order systems. It defines a projection from state space of higher dimension to the state space of lower dimension. To present the utility of the proposed technique, 2 examples are presented with simulation results.