MULTIRATE OUTPUT REGULATION VIA INTERNAL MODEL PRINCIPLE

Abstract

The classical regulator problem is posed for a linear, time invariant, finite-dimensional systems with deterministic disturbance and reference signals in a sampled data framework. The internal model principle is the necessity of this control structure for designing a controller that regulates the output in the continuous feedback case. However, in a multirate sampled-data setup, the output regulation is solvable by designing a controller whose order is less than that of its continuous counterpart. It is shown that the controller becomes the order of exogenous signal generator under standard assumptions on (discrete-time) plant dynamics. When the well known stronger assumptions are used, the controller becomes a static one where the state vector and exogenous signal are estimated at each time step by sampling the output at a faster rate. In this case, the proposed controller takes the form of a full-information based regulator.