Indirect adaptive control for systems with an unknown dead zone

Dead-zone inverse methods have been used in adaptive control schemes to compensate for systems with an unknown dead zone. The problem with these techniques is that steady state error may still exist. It is shown in this paper that a controller with integral action can be used to remove the steady state error arising from an unknown dead zone. By treating the effect of an unknown dead zone as a bounded disturbance being injected into the system, a plant parametrization that is linear in a set of unknown parameters is developed and the estimation algorithm is proposed. A novel feature of the adaptive controller proposed here is the integrating action in the controller. Stability analysis shows that the adaptive scheme ensures boundedness of all closed-loop signals and eliminates tracking errors. As illustrated in a simulation example, the proposed adaptive controller is simple to implement and accurate tracking can be achieved.