Optimal digital redesign of continuous-time controllers (Q1179373)

The paper deals with the problem of designing a digital controller for a continuous dynamic system, assuming that a satisfactory continuous state- feedback control law for the same system exists. The design is based on minimizing an infinite horizon performance index, an integral of a quadratic form expressing the difference between the state of the system controlled respectively by the existing continuous and the new digital controllers. Note, that it is not the standard optimal state tracking problem. By a set of non-trivial manipulations the continuous linear quadratic optimization problem is transformed to a standard discrete-time optimal regulator problem for which the optimal controller can be obtained. The control law consists of three terms based on the actual (measured) state of the digitally controlled dynamic system, the simulated states of the continuously controlled system and the reference input, each with its own gain matrix. The paper presents a numerical example which shows that the digitally controlled output is very close to the original (continuously controlled) one even for a rather large sampling period.