Stability and accuracy analysis of a discrete model reference adaptive controller without and with time delay

Adaptive control techniques can be applied to dynamical systems whose parameters are unknown. We propose a technique based on control and numerical analysis approaches to the study of the stability and accuracy of adaptive control algorithms affected by time delay. In particular, we consider the adaptive minimal control synthesis (MCS) algorithm applied to linear time-invariant plants, due to which, the whole controlled system generated from state and control equations discretized by the zero-order-hold (ZOH) sampling is nonlinear. Hence, we propose two linearization procedures for it: the first is via what we term as physical insight and the second is via Taylor series expansion. The physical insight scheme results in useful methods for a priori selection of the controller parameters and of the discrete-time step. As there is an inherent sampling delay in the process, a fixed one-step delay in the discrete-time MCS controller is introduced. This results in a reduction of both the absolute stability regions and the controller performance. Owing to the shortcomings of ZOH sampling in coping with high-frequency disturbances, a linearly implicit L-stable integrator is also used within a two degree-of-freedom controlled system. The effectiveness of the methodology is confirmed both by simulations and by experimental tests.