A new robust MPC using an approximate convex hull. (June 2018)

• Record Type: Journal Article
• Authors: Sartipizadeh, Hossein; Vincent, Tyrone L.
• Is Part Of: Automatica. Volume 92(2018)
• Page Start: 115
• Page End: 122
• Publication Date: 2018-06
• Subjects: Robust model predictive control -- Approximate convex hull -- Parametric uncertainty -- Measurement uncertainty -- Off-line sampling; Automatic control -- Periodicals; Automation -- Periodicals; 629.805
• DOI: 10.1016/j.automatica.2018.03.010
• Languages: English
• ISSNs: 0005-1098

Abstract

Abstract: A robust Model Predictive Control (MPC) method with a new model uncertainty characterization technique is studied for linear systems. The goal is to reduce the computational complexity of robust MPC. The main element of this method is to trade off the complexity of the representation of the uncertainty set with its accuracy, while still guaranteeing that input, state, and output constraints are met. The mechanism for the trade off is the computation of an approximate convex hull of the model set. This method is an effective tool to eliminate redundancy in the model set and decrease the number of extreme models in robust MPC approaches, resulting in decreased computational complexity. Simulation results illustrate the proposed method.