MODEL-REFERENCE ADAPTIVE-CONTROL WITH L(1+ALPHA) TRACKING

A redesign of model reference adaptive control using a non-quadratic L yapunov function for deriving a new adaptive law leads to a tracking e rror e(t) belonging to the L(1+alpha) signal space for plants with rel ative degree one, where alpha < 0 is a design parameter. A choice of 0 < alpha < 1 leads to the energy of \e(t)\((1+alpha)/2) smaller than t hat with the standard adaptive law for alpha = 1. Simulation results i ndicate that the new adaptive control scheme with 0 < alpha < 1 improv es system tracking performance. Preliminary analysis shows that the ne w scheme with a projection modification is also robust with respect to external disturbances.