Discrete-Time Inverse Optimal Control

Timothy L. Molloy*, Jairo Inga Charaja, Sören Hohmann, Tristan Perez

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

Abstract

In this chapter, we investigate inverse optimal control problems for discrete-time dynamical systems. We pose two discrete-time inverse optimal control problems involving the computation of the parameters of discrete-time optimal control cost functions from data. The problems differ in whether the available data consists of whole or truncated state and control sequences. We present and discuss methods for solving these problems based on bilevel optimization and discrete-time versions of the minimum principle. We specifically show that minimum-principle methods reduce to solving systems of linear equations or quadratic programs under a linear parameterization of the class of cost functions, and admit conditions under which they are guaranteed to provide unique cost-function parameters. Finally, we develop a bespoke technique for solving discrete-time inverse optimal control problems with linear dynamical systems and infinite-horizon quadratic cost functions.

Original languageEnglish
Title of host publicationCommunications and Control Engineering
PublisherSpringer Science and Business Media Deutschland GmbH
Pages41-96
Number of pages56
DOIs
Publication statusPublished - 2022
Externally publishedYes

Publication series

NameCommunications and Control Engineering
ISSN (Print)0178-5354
ISSN (Electronic)2197-7119

Fingerprint

Dive into the research topics of 'Discrete-Time Inverse Optimal Control'. Together they form a unique fingerprint.

Cite this