Sequential approximation of parameter sets for identification with parametric and nonparametric uncertainty

In this paper the problem of approximation of the feasible parameter set for robust identification of a system with parametric and nonparametric uncertainty is considered. A set membership setting is considered where the overall system model is given by a parametric model linear in the unknown parameters and a nonparametric part defined in terms of l₁ or H_∞ norm bounds. A recursive procedure providing an approximation of the parameter set of interest through parallelotopes is presented and an efficient algorithm is proposed. Its computational complexity compares favourably with the more commonly used ellipsoidic approximation schemes. Preliminary numerical results are also reported on some simulation experiments conducted in order to assess the performance of the proposed algorithm