COMPARISON AND EXTENSION OF A DIRECT MODEL-REFERENCE ADAPTIVE-CONTROL PROCEDURE

This paper analyses and extends an easily implemented direct model reference adaptive control procedure. The paper focuses on the major limitation of this control approach which is the satisfaction of a stringent strictly positive real sufficiency condition in order to guarantee asymptotic tracking. Attempts, to date, to address this problem have been unable to relax simultaneously the stringent condition and maintain asymptotic tracking capabilities. This paper presents three different modifications to existing versions of this algorithm which substantially relax the stringent sufficiency condition while providing asymptotic tracking. These three modifications achieve this goal by imposing slight adjustments to existing sufficiency conditions. A simulation example demonstrates that the modifications eliminate the steady state error inherent in the existing methods.