Predictive Control with Probability Constraints for Linear Stochastic Systems

Jiwei Li,Dewei Li,Yugeng Xi
DOI: https://doi.org/10.1049/iet-cta.2016.0915
IF: 2.67
2017-01-01
IET Control Theory and Applications
Abstract:This study develops stochastic model predictive control that guarantees the recursive feasibility of the closed-loop H-infinity performance. Multi-step sets with scaling parameters are proposed to contain the uncertain system trajectory. Meanwhile, by extending the probabilistic invariance to disturbed stochastic systems, we formulate probabilistic constraints as linear matrix inequalities. We show that the introduced scaling parameters enhance the feasibility of H-infinity predictive control and reduce the conservatism of the constraint satisfaction. The designed control algorithm is recursively feasible and stabilises the system in the mean-square sense. A simplified algorithm further reduces much computational burden and makes the proposed approach more practical.
What problem does this paper attempt to address?