A parameter estimation algorithm for nonlinear multivariable systems subject to bounded disturbances

Abstract

In this note, a robust recursive algorithm is derived for identification of nonlinear multivariable systems with unknown by bounded disturbances. Particular emphasis is given to the design of a weighing matrix that ensures consistency of the estimated parameters with the input-output data and the noise constraints, and improves convergence properties. Sufficient conditions for local asymptotic convergence of the algorithm are established. The effectiveness of the proposed algorithm is demonstrated through a numerical example.