T. Fujinaka and S. Omatu (Japan)
LQG Regulators, Pole Shifting, Modal Decomposition, Gain Margin
The design of linear quadratic Gaussian (LQG) regulators with prescribed closed-loop poles is considered. Instead of specifying the weighting matrices of the performance index for an optimal regulator and the covariance matrices for a Kalman filter as in a standard design method, we propose a design method in which the location of the closed-loop poles are specified as the design parameters. The weighting matrices and the covariance matrices which correspond to the prescribed closed-loop poles are constructed, and then the optimal feedback gain and optimal observer gain are calculated. This is an extension to the modal design of LQ regulators which has already been proposed. It is known that the infinite gain margin property of LQ regulators does not extend to LQG regulators, and the gain margin property of the proposed LQG regulators is discussed.
Important Links:
Go Back
