Learning strategies for wet clutch control

Gregory Pinte; J. Stoev; Wim Symens; Abhishek Dutta; Yu Zhong; Bart Wyns; R. De Keyser; Bruno Depraetere; Jan Swevers; Matteo Gagliolo; Ann Nowe

Learning strategies for wet clutch control

Gregory Pinte, J. Stoev, Wim Symens, Abhishek Dutta, Yu Zhong, Bart Wyns, R. De Keyser, Bruno Depraetere, Jan Swevers, Matteo Gagliolo, Ann Nowe

Informatics and Applied Informatics

Research output: Chapter in Book/Report/Conference proceeding › Conference paper

5 Citations (Scopus)

Abstract

This paper presents an overview of model-based
(Iterative Learning Control, Model Predictive Control and
Iterative Optimization) and non-model-based (Genetic-based
Machine Learning and Reinforcement Learning) learning
strategies for the control of wet clutches. Based on theoretical
considerations and a validation on an experimental test bench
containing wet clutches, the benefits and drawbacks of the
different strategies are compared. Although after convergence
a good engagement quality can be obtained by all strategies,
only model-based strategies are suited for online applicability.
The convergence time for non-model-based strategies is too long
such that they can only be applied during an offline calibration
phase.

Original language	English
Title of host publication	15th International Conference on System Theory, Control and Computing - ICSTCC 2011
Editors	Mihail Voicu
Publisher	IEEE
Pages	467-474
Number of pages	8
ISBN (Print)	978-973-621-322-9
Publication status	Published - 14 Oct 2011
Event	International Conference on System Theory, Control and Computing - Duration: 14 Nov 2011 → …

Publication series

Name	15th International Conference on System Theory, Control and Computing - ICSTCC 2011

Conference

Conference	International Conference on System Theory, Control and Computing
Abbreviated title	ICSTCC 2011
Period	14/11/11 → …

Bibliographical note

Mihail Voicu

Keywords

control
wet clutch
genetic algorithms
reinforcement learning
iterative learning control
model predictive control

Handle.net

Persistent link

Cite this

Pinte, G., Stoev, J., Symens, W., Dutta, A., Zhong, Y., Wyns, B., De Keyser, R., Depraetere, B., Swevers, J., Gagliolo, M., & Nowe, A. (2011). Learning strategies for wet clutch control. In M. Voicu (Ed.), 15th International Conference on System Theory, Control and Computing - ICSTCC 2011 (pp. 467-474). (15th International Conference on System Theory, Control and Computing - ICSTCC 2011). IEEE.

@inproceedings{b0b1a04d1af44689923d3e50ef1907d5,

title = "Learning strategies for wet clutch control",

abstract = "This paper presents an overview of model-based (Iterative Learning Control, Model Predictive Control and Iterative Optimization) and non-model-based (Genetic-based Machine Learning and Reinforcement Learning) learning strategies for the control of wet clutches. Based on theoretical considerations and a validation on an experimental test bench containing wet clutches, the benefits and drawbacks of the different strategies are compared. Although after convergence a good engagement quality can be obtained by all strategies, only model-based strategies are suited for online applicability. The convergence time for non-model-based strategies is too long such that they can only be applied during an offline calibration phase.",

keywords = "control, wet clutch, genetic algorithms, reinforcement learning, iterative learning control, model predictive control",

author = "Gregory Pinte and J. Stoev and Wim Symens and Abhishek Dutta and Yu Zhong and Bart Wyns and {De Keyser}, R. and Bruno Depraetere and Jan Swevers and Matteo Gagliolo and Ann Nowe",

note = "Mihail Voicu; International Conference on System Theory, Control and Computing <br/><br/>, ICSTCC 2011 ; Conference date: 14-11-2011",

year = "2011",

month = oct,

day = "14",

language = "English",

isbn = "978-973-621-322-9",

series = "15th International Conference on System Theory, Control and Computing - ICSTCC 2011",

publisher = "IEEE",

pages = "467--474",

editor = "Mihail Voicu",

booktitle = "15th International Conference on System Theory, Control and Computing - ICSTCC 2011",

}

Pinte, G, Stoev, J, Symens, W, Dutta, A, Zhong, Y, Wyns, B, De Keyser, R, Depraetere, B, Swevers, J, Gagliolo, M & Nowe, A 2011, Learning strategies for wet clutch control. in M Voicu (ed.), 15th International Conference on System Theory, Control and Computing - ICSTCC 2011. 15th International Conference on System Theory, Control and Computing - ICSTCC 2011, IEEE, pp. 467-474, International Conference on System Theory, Control and Computing

, 14/11/11.

Learning strategies for wet clutch control. / Pinte, Gregory; Stoev, J.; Symens, Wim et al.
15th International Conference on System Theory, Control and Computing - ICSTCC 2011. ed. / Mihail Voicu. IEEE, 2011. p. 467-474 (15th International Conference on System Theory, Control and Computing - ICSTCC 2011).

Research output: Chapter in Book/Report/Conference proceeding › Conference paper

TY - GEN

T1 - Learning strategies for wet clutch control

AU - Pinte, Gregory

AU - Stoev, J.

AU - Symens, Wim

AU - Dutta, Abhishek

AU - Zhong, Yu

AU - Wyns, Bart

AU - De Keyser, R.

AU - Depraetere, Bruno

AU - Swevers, Jan

AU - Gagliolo, Matteo

AU - Nowe, Ann

N1 - Mihail Voicu

PY - 2011/10/14

Y1 - 2011/10/14

N2 - This paper presents an overview of model-based (Iterative Learning Control, Model Predictive Control and Iterative Optimization) and non-model-based (Genetic-based Machine Learning and Reinforcement Learning) learning strategies for the control of wet clutches. Based on theoretical considerations and a validation on an experimental test bench containing wet clutches, the benefits and drawbacks of the different strategies are compared. Although after convergence a good engagement quality can be obtained by all strategies, only model-based strategies are suited for online applicability. The convergence time for non-model-based strategies is too long such that they can only be applied during an offline calibration phase.

AB - This paper presents an overview of model-based (Iterative Learning Control, Model Predictive Control and Iterative Optimization) and non-model-based (Genetic-based Machine Learning and Reinforcement Learning) learning strategies for the control of wet clutches. Based on theoretical considerations and a validation on an experimental test bench containing wet clutches, the benefits and drawbacks of the different strategies are compared. Although after convergence a good engagement quality can be obtained by all strategies, only model-based strategies are suited for online applicability. The convergence time for non-model-based strategies is too long such that they can only be applied during an offline calibration phase.

KW - control

KW - wet clutch

KW - genetic algorithms

KW - reinforcement learning

KW - iterative learning control

KW - model predictive control

M3 - Conference paper

SN - 978-973-621-322-9

T3 - 15th International Conference on System Theory, Control and Computing - ICSTCC 2011

SP - 467

EP - 474

BT - 15th International Conference on System Theory, Control and Computing - ICSTCC 2011

A2 - Voicu, Mihail

PB - IEEE

T2 - International Conference on System Theory, Control and Computing <br/><br/>

Y2 - 14 November 2011

ER -

Learning strategies for wet clutch control

Abstract

Publication series

Conference

Bibliographical note

Keywords

Handle.net

Fingerprint

Cite this