Data-driven neural feedforward controller design for industrial linear motors

Yuk Hang Yuen, Mircea Lazar, Hans Butler

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Downloads (Pure)

Abstract

In this paper we consider the problem of feedforward controller design for industrial linear motors. These motors are safety-critical high-precision mechatronics systems that pose stringent requirements on the feedforward design: safe and predictable behavior for the desired motion profiles, tracking performance within the 10μ m range in the presence of nonlinear friction and real-time implementation within the 1ms range. We investigate and compare several possibilities to design data-driven feedforward controllers using neural networks (NN) and we show that a two-step inverse estimation method is the most suitable approach, due to robustness to noisy data. We also show that basic knowledge about the system dynamics and the friction behavior can be exploited to design neural feedforward controllers with a simple structure, suitable for real-time implementation in industrial linear motors. The developed data-driven neural feedforward controllers are tested and compared with standard mass-acceleration feedforward and iterative learning controllers in realistic simulations.

Original languageEnglish
Title of host publication2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings
EditorsRadu-Emil Precup
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Pages461-467
Number of pages7
ISBN (Electronic)9781728106991
DOIs
Publication statusPublished - Oct 2019
Event23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Sinaia, Romania
Duration: 9 Oct 201911 Oct 2019

Publication series

Name2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings

Conference

Conference23rd International Conference on System Theory, Control and Computing, ICSTCC 2019
CountryRomania
CitySinaia
Period9/10/1911/10/19

Fingerprint

Linear Motor
Linear motors
Feedforward
Data-driven
Controller Design
Controllers
Controller
Friction
Mechatronics
Real-time
Noisy Data
Dynamical systems
System Dynamics
Range of data
Neural networks
Safety
Neural Networks
Robustness
Motion
Requirements

Keywords

  • Data-driven control
  • Feedforward control
  • Linear motors
  • Neural networks

Cite this

Yuen, Y. H., Lazar, M., & Butler, H. (2019). Data-driven neural feedforward controller design for industrial linear motors. In R-E. Precup (Ed.), 2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings (pp. 461-467). [8885434] (2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings). Piscataway: Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICSTCC.2019.8885434
Yuen, Yuk Hang ; Lazar, Mircea ; Butler, Hans. / Data-driven neural feedforward controller design for industrial linear motors. 2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings. editor / Radu-Emil Precup. Piscataway : Institute of Electrical and Electronics Engineers, 2019. pp. 461-467 (2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings).
@inproceedings{bd9ab1abbd0d460798d247679c7aa6a5,
title = "Data-driven neural feedforward controller design for industrial linear motors",
abstract = "In this paper we consider the problem of feedforward controller design for industrial linear motors. These motors are safety-critical high-precision mechatronics systems that pose stringent requirements on the feedforward design: safe and predictable behavior for the desired motion profiles, tracking performance within the 10μ m range in the presence of nonlinear friction and real-time implementation within the 1ms range. We investigate and compare several possibilities to design data-driven feedforward controllers using neural networks (NN) and we show that a two-step inverse estimation method is the most suitable approach, due to robustness to noisy data. We also show that basic knowledge about the system dynamics and the friction behavior can be exploited to design neural feedforward controllers with a simple structure, suitable for real-time implementation in industrial linear motors. The developed data-driven neural feedforward controllers are tested and compared with standard mass-acceleration feedforward and iterative learning controllers in realistic simulations.",
keywords = "Data-driven control, Feedforward control, Linear motors, Neural networks",
author = "Yuen, {Yuk Hang} and Mircea Lazar and Hans Butler",
year = "2019",
month = "10",
doi = "10.1109/ICSTCC.2019.8885434",
language = "English",
series = "2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers",
pages = "461--467",
editor = "Radu-Emil Precup",
booktitle = "2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings",
address = "United States",

}

Yuen, YH, Lazar, M & Butler, H 2019, Data-driven neural feedforward controller design for industrial linear motors. in R-E Precup (ed.), 2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings., 8885434, 2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings, Institute of Electrical and Electronics Engineers, Piscataway, pp. 461-467, 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019, Sinaia, Romania, 9/10/19. https://doi.org/10.1109/ICSTCC.2019.8885434

Data-driven neural feedforward controller design for industrial linear motors. / Yuen, Yuk Hang; Lazar, Mircea; Butler, Hans.

2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings. ed. / Radu-Emil Precup. Piscataway : Institute of Electrical and Electronics Engineers, 2019. p. 461-467 8885434 (2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings).

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Data-driven neural feedforward controller design for industrial linear motors

AU - Yuen, Yuk Hang

AU - Lazar, Mircea

AU - Butler, Hans

PY - 2019/10

Y1 - 2019/10

N2 - In this paper we consider the problem of feedforward controller design for industrial linear motors. These motors are safety-critical high-precision mechatronics systems that pose stringent requirements on the feedforward design: safe and predictable behavior for the desired motion profiles, tracking performance within the 10μ m range in the presence of nonlinear friction and real-time implementation within the 1ms range. We investigate and compare several possibilities to design data-driven feedforward controllers using neural networks (NN) and we show that a two-step inverse estimation method is the most suitable approach, due to robustness to noisy data. We also show that basic knowledge about the system dynamics and the friction behavior can be exploited to design neural feedforward controllers with a simple structure, suitable for real-time implementation in industrial linear motors. The developed data-driven neural feedforward controllers are tested and compared with standard mass-acceleration feedforward and iterative learning controllers in realistic simulations.

AB - In this paper we consider the problem of feedforward controller design for industrial linear motors. These motors are safety-critical high-precision mechatronics systems that pose stringent requirements on the feedforward design: safe and predictable behavior for the desired motion profiles, tracking performance within the 10μ m range in the presence of nonlinear friction and real-time implementation within the 1ms range. We investigate and compare several possibilities to design data-driven feedforward controllers using neural networks (NN) and we show that a two-step inverse estimation method is the most suitable approach, due to robustness to noisy data. We also show that basic knowledge about the system dynamics and the friction behavior can be exploited to design neural feedforward controllers with a simple structure, suitable for real-time implementation in industrial linear motors. The developed data-driven neural feedforward controllers are tested and compared with standard mass-acceleration feedforward and iterative learning controllers in realistic simulations.

KW - Data-driven control

KW - Feedforward control

KW - Linear motors

KW - Neural networks

UR - http://www.scopus.com/inward/record.url?scp=85075034128&partnerID=8YFLogxK

U2 - 10.1109/ICSTCC.2019.8885434

DO - 10.1109/ICSTCC.2019.8885434

M3 - Conference contribution

AN - SCOPUS:85075034128

T3 - 2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings

SP - 461

EP - 467

BT - 2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings

A2 - Precup, Radu-Emil

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

ER -

Yuen YH, Lazar M, Butler H. Data-driven neural feedforward controller design for industrial linear motors. In Precup R-E, editor, 2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings. Piscataway: Institute of Electrical and Electronics Engineers. 2019. p. 461-467. 8885434. (2019 23rd International Conference on System Theory, Control and Computing, ICSTCC 2019 - Proceedings). https://doi.org/10.1109/ICSTCC.2019.8885434