Memory reduction of rate-dependent Prandtl–Ishlinskii compensators in applications on high-precision motion systems

Abstract

Inverse compensation of rate-dependent Prandtl–Ishlinskii operators with finitely many thresholds can be carried out explicitly under some structural conditions. However, if the number of thresholds is very high, the formulas become complicated. A simple and elegant framework for the investigation of such problems encompasses a recently proposed extension of the class of rate-dependent Prandtl–Ishlinskii operators to the case of a whole continuum of play operators with time-dependent thresholds. Indeed, such a theory allows for reducing the number of necessary thresholds in the compensation procedure and estimating the errors of the memory-discrete compensation in terms of the distance between the thresholds and weights of the individual plays. Following these results, our goal in this work is the validation of these theoretical models via numerical simulations and experimental results. In particular, we show that high accuracy of the hysteresis compensation algorithm can be achieved even with a relatively small number of thresholds.