Optimal Decoupling Control

Abstract

A new approach to optimal decoupling for multi-input multi-output (MIMO), linear, time-invariant control systems, is presented. This general optimal decoupling method is based on the minimization of componentwise performance indices. These indices can be obtained from alternative componentwise mathematical representations of the original system. Furthermore, using these representations, optimal decoupling problems can be written in terms of ordinary optimal control problems. Then one can use classical optimal control techniques such as LQR of Hto obtain the ideal solutions to the optimal decoupling problems.<P>However these techniques can not be used to obtain practical closed-loop solutions. In this dissertation, fixed structure control techniques are used to derive the optimal and sub-optimal state feedback control law. For this purpose, a globally convergent numerical algorithm is derived. In addition some other global sub-optimal solutions are studied. This approach is used in the design of an attitude flight control of the UH-60A helicopter in hover.