Indirect-feedback Stochastic MPC for tracking piecewise constant references

Abstract

This paper presents an indirect-feedback stochastic model predictive control (IF-SMPC) algorithm for tracking piecewise constant references. The IF-SMPC strategy for regulation is combined with the artificial reference approach to reach a neighborhood of the optimal admissible piecewise constant reference. The main objectives of the proposed strategy are (i) to avoid feasibility loss due to target change, (ii) to ensure the convergence to the neighborhood of the optimal admissible target, and (iii) to guarantee the prescribed individual chance constraint satisfaction provided by an indirect-feedback SMPC solution. It is shown the proposed IF-SMPC may provide a potentially enlarged feasible region if compared with the robust SMPC alternative. The formal analyses of the desired properties are presented. A case study based on a DC-DC converter benchmark is presented to illustrate the usefulness of the proposed IF-SMPC for tracking piecewise constant references.