Twisting algorithm adaptation for control of electropneumatic actuators

The paper proposes a new adaptive version of the twisting algorithm, which is a traditional second order sliding mode controller. Due to dynamic adaptation of the gains, the controller design does not require complete information on the uncertainties/perturbations (bounds). It allows to decrease the gain magnitude and then the oscillations caused by a too large control magnitude: it improves the performance of the closed-loop system (accuracy). To show the feasibility of the approach, the methodology is applied to position control of an electropneumatic actuator.