Robust stability constraints for predictive control

In this paper new conditions are presented which guarantee robust global asymptotic stability of non-linear controllers for open-loop stable non-linear systems with additive l₁-norm bounded model uncertainty. The resulting robust stability constraints are control method independent, linear and very unconservative because they are updated at every sample time. Based on these new conditions a predictive control method is designed for stable linear models with additive model uncertainty. Robust global asymptotic stability and offset free tracking are guaranteed for asymptotically constant reference trajectories and disturbances.