Recursive \(l^1\) robust identification algorithms (Q2725415)

The authors consider the problem of \(l^1\) robust identification for dynamic systems with norm bounded, unstructured uncertainty. They first show that the set of considered model parameters is a convex polyhedron, assuming less a priori system knowledge and bounded noise. Based on the above result, they propose a simple, recursive algorithm for determining the nominal model and the identification error bound. Moreover, they analyze the convergence of the proposed recursive algorithm. A simulation example is presented to illustrate its effectiveness.