Markov Parameters Sequence Identification Oriented Data-Driven LQ=H∞ Robust Preview Control

Abstract

In this paper, the data-driven robust preview control problem is addressed based on Markov parameters sequence identification and augmented modelling technique. The involved analysis and synthesis are composed of three parts. First, data-based state-space model is established by augmenting input/output data, finite window previewable signals and tracking errors. Then, the Markov parameters sequence is identified, which enables the determination of data model matrices. In the following, the mixed linear quadratic (LQ) and H∞ criterions are used to optimize the robust preview control gains, and the specified preview control policy containing data feedback control, integral operation and preview action is finally obtained. The application to injection velocity control of injection molding process verifies the effectiveness of proposed results.