Data-driven sliding mode control: a new approach based on optimization

Abstract

ABSTRACT In this paper a model free sliding mode controller (SMC) based on a novel reaching law for SISO nonlinear systems is provided. The novelty of the work lies in the approach used to achieve the reaching law. The proposed reaching law is more general than those proposed in recent works. Naturally, it is consistent with the proposed control law, which is composed of an optimal term and an exponential switching term. A proper cost function forces the sliding variable and its derivative to vanish. The switching part consists of an exponential term, which decreases the quasi-sliding domain. It is shown analytically that the sliding function enters and remains in the quasi-sliding bound. In addition, the UUB stability of error is rigorously proved. Computer simulation in different scenarios and comparison to the simple model free SMC and model free adaptive control methods has been provided to show better performance of the proposed method, compared with the mentioned methods.