A GENERAL RESULT ON THE STABILIZATION OF LINEAR-SYSTEMS USING BOUNDEDCONTROLS

We present two constructions of controllers that globally stabilize li near systems subject to control saturation. We allow essentially arbit rary saturation functions. The only conditions imposed on the system a re the obvious necessary ones, namely that no eigenvalues of the uncon trolled system have positive real part and that the standard stabiliza bility rank condition hold. One of the constructions is in terms of a ''neural-network type'' one-hidden layer architecture, while the other one is in terms of cascades of linear maps and saturations.