MODEL-REFERENCE ROBUST-CONTROL OF A CLASS OF SISO SYSTEMS

A new control design technique, model reference robust control (MRRC), is introduced for a class of SISO systems which contain unknown param eters, possible nonlinear uncertainties, and additive bounded disturba nces. The design methodology is a natural, nontrivial extension of mod el reference adaptive control (MRAC) which is essential to achieving r obust stability and performance for linear time-invariant systems. The methodology also represents an important step toward achieving robust stability for time-varying and nonlinear systems. MRRC requires only input and output measurements of the system, rather than the full stat e feedback and structural conditions on uncertainties required by exis ting robust control results. MRRC is developed from existing model ref erence control (MRC) in a manner similar to MRAC. An intermediate resu lt gives conditions under which MRRC yields exponentially asymptotic s tability. The general result yielding uniformly ultimately bounded sta bility is then developed. A scalar example provides intuition into why the control works against a wide class of uncertainties and reveals t he implicit learning capability of MRRC.