Gaussian networks for control of a class of systems with friction

Identification and stable adaptive control of a class of systems with friction is considered using Gaussian networks. Preliminary results are presented for a proposed strategy using an experimental system consisting of a DC motor and a load. The nonlinearity due to friction, which is significant at low velocities, is first identified using a Gaussian network and then compensated for using the network in a feedforward mode. The Gaussian network is shown to have a 'general' structure suited for friction problems. The network development takes advantage of the constructive methodology for generating stable adaptive laws for Gaussian networks proposed by Sanner and Slotine (1993).