$\mathcal{N}$IPM-MPC: An Efficient Null-Space Method Based Interior-Point Method for Model Predictive Control

TL;DR

This paper introduces a novel sparse null-space basis for interior-point methods in linear MPC, significantly reducing computational effort and outperforming existing solvers in speed.

Contribution

It proposes a sparse null-space basis based on virtual controls that preserves block structure and reduces factorizations in interior-point MPC algorithms.

Findings

Outperforms existing solvers in computational speed.

Reduces the number of matrix factorizations per iteration.

Preserves block diagonal structure for efficiency.

Abstract

Linear Model Predictive Control (MPC) is a widely used method to control systems with linear dynamics. Efficient interior-point methods have been proposed which leverage the block diagonal structure of the quadratic program (QP) resulting from the receding horizon control formulation. However, they require two matrix factorizations per interior-point method iteration, one each for the computation of the dual and the primal. Recently though an interior point method based on the null-space method has been proposed which requires only a single decomposition per iteration. While the then used null-space basis leads to dense null-space projections, in this work we propose a sparse null-space basis which preserves the block diagonal structure of the MPC matrices. Since it is based on the inverse of the transfer matrix we introduce the notion of so-called virtual controls which enables just that invertibility. A combination of the reduced number of factorizations and omission of the evaluation of the dual lets our solver outperform others in terms of computational speed by an increasing margin dependent on the number of state and control variables.