Discrete-time multivariable MRAC schemes with output sensor uncertainty compensation

This paper studies model reference adaptive control (MRAC) of discrete-time multi-input multi-output systems in the presence of output sensor uncertainties. Two multivariable MRAC schemes are developed, one for systems with known dynamics and one for systems with unknown dynamics, both with an adaptive compensation design for dealing with parametrizable sensor uncertainties when reliable output measurements are not available for feedback. Complete system and control parametrizations are derived for adaptive compensation designs; in particular, for the unknown dynamics case, a new control parametrization is employed to reach a desirable linear model for adaptation. Stability and convergence analysis is given. Simulation results on a discretized MIMO aircraft lateral model are presented to show that adaptive sensor compensation can significantly improve system tracking performance, for parametrizable sensor uncertainties.