Transient Performance Improvement in Continuous-time Model Reference Adaptive Control: An L1 Formulation

This paper presents an L1 formulation of the transient performance issue in model reference adaptive control. Specifically, it is shown that, in the case of known high frequency gain, the parametric uncertainty can be viewed as an input disturbance. Unlike as in standard L1 optimal control, this disturbance cannot be apriori assumed to be bounded. Nevertheless, an L1 optimal controller for the rejection of persistent bounded disturbances is designed, and is shown to be identical to a modified model refererence adaptive control scheme, proposed earlier for achieving performance improvement.