Adaptive Observer for Regression Models with External Time–Dependent Disturbances

Abstract

In this paper, a robust adaptive observer is proposed for dynamical disturbed regression models. For the constant unknown parameters case, the proposed algorithm ensures asymptotic convergence to zero of the parameter identification error in presence of time-dependent external disturbances. For the case of the time-varying parameters, the parameter identification error converges asymptotically to an arbitrarily small region around the origin in presence of time-dependent external disturbances. The synthesis of the adaptive observer is based on the solution of a linear matrix inequality. Numerical simulations illustrate the convergence properties of the proposed algorithm.