Tracking control of mechanical systems in the presence of nonlinear dynamic friction effects

We design an observer-based, exact model knowledge position tracking controller for a second-order mechanical system with nonlinear load dynamics and the nonlinear, dynamic friction model proposed in Canudas de Wit et al. (1995). Since the controller requires an estimate of the immeasurable friction state, we demonstrate how the friction dynamics can be exploited to design three different observers which foster different transient response characteristics for the composite closed-loop system. We then present two adaptive controllers which utilize nonlinear observer/filter structures to provide for asymptotic position tracking while compensating for selected parametric uncertainty.