Convex reformulations for a special class of nonlinear MPC problems

TL;DR

This paper introduces a method to solve a specific class of nonlinear model predictive control problems by reformulating them into a finite set of convex subproblems using exact linearization, simplifying the solution process.

Contribution

It presents a novel convex reformulation approach for a special class of nonlinear MPC problems leveraging input-state linearization.

Findings

Reformulation into convex subproblems is exact for the considered system class.

The approach simplifies solving NMPC problems by decomposing them into convex parts.

Numerical examples demonstrate the effectiveness of the method.

Abstract

We show how the solution to NMPC problems for a special type of input-affine discrete-time systems can be obtained by reformulating the underlying non-convex optimal control problem in terms of a finite number of convex subproblems. The reformulation is facilitated by exact (input-state) linearization, which is shown to provide beneficial properties for the treated class of systems. We characterize possible types of the resulting convex subproblems and illustrate our approach with three numerical examples.