A Tustin-Based Discrete-Time Implementation of the Generalized Clegg Integrator

Abstract

This paper proposes a novel discrete-time (DT) implementation of the generalized Clegg integrator (GCI), which is an integrator that resets its state to a fraction of the original state when its input is equal to zero. The implementation is derived by discretizing a continuous-time (CT) GCI using the Tustin discretization method. By means of a numerical validation it is shown that the state of the DT GCI is identical to its CT counterpart when both are subject to an input which is linearly interpolated between samples, as expected when using this discretization method. For a general CT input which is not linearly interpolated between samples, a numerical comparison is made between the state of the novel DT GCI and the CT GCI. At samples with linear behaviour, the state mismatch is equivalent to the one observed between their linear counterparts. At samples with resetting behaviour, the mismatch even reduces compared to previous samples, as a consequence of (partially) resetting the state mismatch.