Feasible Controller Design for Stochastic Systems
Constantine Frangos, Y. Yavin · Journal of Guidance Control and Dynamics · 1997
Acontroldesignmethodispresented forlinearplantmodelswith stochasticstatedisturbancesandmeasurement noise such that time domain and frequency domain specie cations are met. A time invariant linear quadratic Gaussian (LQG) controller structure is employed. The LQG controller is parametrized by the diagonal elements ofthestateweighting matrix Qand statenoiseintensitymatrix L.Astandard searchalgorithmisappliediteratively to compute Q and L such that all of the specie cations are satise ed. At each iteration step of the search algorithm, an LQG problem is solved.