Generalization of L1 adaptive control architecture for switching estimation laws

This paper presents a modification of the estimation laws in a typical L₁ adaptive controller, by replacing the projection-based integrator in the estimation loop by a switching law with a dead-zone. We consider a class of linear systems with parametric uncertainty and input disturbance. Similar to prior solutions in L₁ adaptive control theory, the feedback structure is comprised of the three main elements, involving the state predictor, the adaptation laws and the lowpass filter. We derive the performance bounds for both the system output and the control signal and show that these can be adjusted by a suitable choice of a dead-zone interval.