Optimal regulation within spatial constraints. An application to flexible structures

The contribution of this research is a candidate controller architecture for application to lightly damped, flexible vehicles, where stiffness and mass are distributed parameters. The architecture acknowledges that a reduced order controller is required. Singular perturbation theory and the spatial shaping of control inputs and observation residuals are then coupled to derive an appropriate, asymptotically correct, low frequency plant representation, and a cascaded estimator/optimal regulator is used to construct a linear feedback law based on this low order model. The primary benefit of this approach is the capability to rigorously construct control laws for systems where the dimensions of the control vector, output vector, and the plant dynamics are high. Design examples are given which illustrate that the reduced order designs compare favorably to the optimal fullstate regulator.