Feedback Linearization of Multilinear Time-invariant Systems using Tensor Decomposition Methods

Kai Kruppa, Gerwald Lichtenberg

2018

Abstract

Multilinear Time-Invariant (MTI) Systems can represent or approximate nonlinear systems behaviour in a tensor framework. The paper derives efficient algorithms for feedback linearization of MTI systems. Basic operations like products and partial derivatives of multilinear or higher order polynomials and Lie derivatives can be defined in terms of Canonical Polyadic (CP) tensors, as well as the parameters of MTI systems. The feedback linearizing controller design algorithm given in the paper results in a controller of a known fixed structure with predictable memory demand, which is a great advantage in case of implementation. The structure does not depend on the MTI plant and has to be adjusted to the application by setting its parameters only. Moreover, the MTI algorithm only involves numerical and no symbolic computations. This is relevant for large scale applications like heterogeneous networks e.g. for heat and power distribution, since the model can be represented by decomposed tensors to reduce the storage demand.

Paper Citation

in Harvard Style

Kruppa K. and Lichtenberg G. (2018). Feedback Linearization of Multilinear Time-invariant Systems using Tensor Decomposition Methods.In Proceedings of 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH, ISBN 978-989-758-323-0, pages 232-243. DOI: 10.5220/0006852802320243

in Bibtex Style

@conference{simultech18,
author={Kai Kruppa and Gerwald Lichtenberg},
title={Feedback Linearization of Multilinear Time-invariant Systems using Tensor Decomposition Methods},
booktitle={Proceedings of 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH,},
year={2018},
pages={232-243},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0006852802320243},
isbn={978-989-758-323-0},
}

in EndNote Style

TY - CONF

JO - Proceedings of 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH,
TI - Feedback Linearization of Multilinear Time-invariant Systems using Tensor Decomposition Methods
SN - 978-989-758-323-0
AU - Kruppa K.
AU - Lichtenberg G.
PY - 2018
SP - 232
EP - 243
DO - 10.5220/0006852802320243