Disturbances

Abstract

In linear quadratic control the objective of the controller is to minimize the effect of the initial condition on the (quadratic) cost functional. However, in many systems, uncontrolled inputs, known as disturbances are present. A controller can reduce the response to the disturbances. Even with the same cost function as for linear quadratic control, the problem statement changes. The problems of a single disturbance, and also of an unknown disturbance are considered. The solutions to these problems lead to H:2 and H control problems respectively. The framework for approximation is described, as are appropriate computation methods. As for other control system objectives, actuator location is part of control system design.