Adaptive Control Learning Based on a Similarity Measure

T. Rocha, S. Paredes, J. Henriques, P. Carvalho, A. Cardoso, P. Gil

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This work proposes a learning methodology to be employed in an adaptive controller strategy. It is based on a similarity approach along with a pole placement technique, by combining ideas from adaptive control and machine learning areas. The learning scheme is propped on the hypothesis that the current characterization of a given system can be achieved from the analysis of past similar behaviors. The main assumption is that data gathered from past experiments, during the operation, can be used on-to line reduce the uncertainty of a model that describes the system. As result, the strategy contributes to improve the performance of a controller based on that model. Two main steps are involved. In the first step a similarity analysis is performed, enabling to find in the historical a set of patterns (input/output time series) similar to the current condition. Then, in a second step, these time series are used to estimate the parameters of a linear model, that are employed afterwards in a pole placement control tuning. The applicability of the proposed approach is assessed on a benchmark nonlinear process, namely a continuous stirred tank reactor (CSTR), showing a better performance than with fixed PI and pole placement controllers
Original languageEnglish
Title of host publication2018 13th APCA International Conference on Control and Soft Computing (CONTROLO)
Pages91-97
Number of pages7
DOIs
Publication statusPublished - 1 Jun 2018
Event13th APCA International Conference on Automatic Control and Soft Computing (CONTROLO) - Ponta Delgada, Portugal
Duration: 4 Jun 20186 Jun 2018

Conference

Conference13th APCA International Conference on Automatic Control and Soft Computing (CONTROLO)
CountryPortugal
City Ponta Delgada
Period4/06/186/06/18

Fingerprint

Poles
Controllers
Time series
Learning systems
Tuning
Experiments
Uncertainty

Keywords

  • adpative control
  • case based reasoning
  • Machine Learning
  • Chemical reactors
  • continuous stirred tank reactor

Cite this

Rocha, T., Paredes, S., Henriques, J., Carvalho, P., Cardoso, A., & Gil, P. (2018). Adaptive Control Learning Based on a Similarity Measure. In 2018 13th APCA International Conference on Control and Soft Computing (CONTROLO) (pp. 91-97) https://doi.org/10.1109/CONTROLO.2018.8514277
Rocha, T. ; Paredes, S. ; Henriques, J. ; Carvalho, P. ; Cardoso, A. ; Gil, P. / Adaptive Control Learning Based on a Similarity Measure. 2018 13th APCA International Conference on Control and Soft Computing (CONTROLO). 2018. pp. 91-97
@inproceedings{768c54b08886433889807dd456fa8fcc,
title = "Adaptive Control Learning Based on a Similarity Measure",
abstract = "This work proposes a learning methodology to be employed in an adaptive controller strategy. It is based on a similarity approach along with a pole placement technique, by combining ideas from adaptive control and machine learning areas. The learning scheme is propped on the hypothesis that the current characterization of a given system can be achieved from the analysis of past similar behaviors. The main assumption is that data gathered from past experiments, during the operation, can be used on-to line reduce the uncertainty of a model that describes the system. As result, the strategy contributes to improve the performance of a controller based on that model. Two main steps are involved. In the first step a similarity analysis is performed, enabling to find in the historical a set of patterns (input/output time series) similar to the current condition. Then, in a second step, these time series are used to estimate the parameters of a linear model, that are employed afterwards in a pole placement control tuning. The applicability of the proposed approach is assessed on a benchmark nonlinear process, namely a continuous stirred tank reactor (CSTR), showing a better performance than with fixed PI and pole placement controllers",
keywords = "adpative control, case based reasoning, Machine Learning, Chemical reactors, continuous stirred tank reactor",
author = "T. Rocha and S. Paredes and J. Henriques and P. Carvalho and A. Cardoso and P. Gil",
year = "2018",
month = "6",
day = "1",
doi = "10.1109/CONTROLO.2018.8514277",
language = "English",
isbn = "78-1-5386-5346-3",
pages = "91--97",
booktitle = "2018 13th APCA International Conference on Control and Soft Computing (CONTROLO)",

}

Rocha, T, Paredes, S, Henriques, J, Carvalho, P, Cardoso, A & Gil, P 2018, Adaptive Control Learning Based on a Similarity Measure. in 2018 13th APCA International Conference on Control and Soft Computing (CONTROLO). pp. 91-97, 13th APCA International Conference on Automatic Control and Soft Computing (CONTROLO), Ponta Delgada, Portugal, 4/06/18. https://doi.org/10.1109/CONTROLO.2018.8514277

Adaptive Control Learning Based on a Similarity Measure. / Rocha, T.; Paredes, S.; Henriques, J.; Carvalho, P.; Cardoso, A.; Gil, P.

2018 13th APCA International Conference on Control and Soft Computing (CONTROLO). 2018. p. 91-97.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Adaptive Control Learning Based on a Similarity Measure

AU - Rocha, T.

AU - Paredes, S.

AU - Henriques, J.

AU - Carvalho, P.

AU - Cardoso, A.

AU - Gil, P.

PY - 2018/6/1

Y1 - 2018/6/1

N2 - This work proposes a learning methodology to be employed in an adaptive controller strategy. It is based on a similarity approach along with a pole placement technique, by combining ideas from adaptive control and machine learning areas. The learning scheme is propped on the hypothesis that the current characterization of a given system can be achieved from the analysis of past similar behaviors. The main assumption is that data gathered from past experiments, during the operation, can be used on-to line reduce the uncertainty of a model that describes the system. As result, the strategy contributes to improve the performance of a controller based on that model. Two main steps are involved. In the first step a similarity analysis is performed, enabling to find in the historical a set of patterns (input/output time series) similar to the current condition. Then, in a second step, these time series are used to estimate the parameters of a linear model, that are employed afterwards in a pole placement control tuning. The applicability of the proposed approach is assessed on a benchmark nonlinear process, namely a continuous stirred tank reactor (CSTR), showing a better performance than with fixed PI and pole placement controllers

AB - This work proposes a learning methodology to be employed in an adaptive controller strategy. It is based on a similarity approach along with a pole placement technique, by combining ideas from adaptive control and machine learning areas. The learning scheme is propped on the hypothesis that the current characterization of a given system can be achieved from the analysis of past similar behaviors. The main assumption is that data gathered from past experiments, during the operation, can be used on-to line reduce the uncertainty of a model that describes the system. As result, the strategy contributes to improve the performance of a controller based on that model. Two main steps are involved. In the first step a similarity analysis is performed, enabling to find in the historical a set of patterns (input/output time series) similar to the current condition. Then, in a second step, these time series are used to estimate the parameters of a linear model, that are employed afterwards in a pole placement control tuning. The applicability of the proposed approach is assessed on a benchmark nonlinear process, namely a continuous stirred tank reactor (CSTR), showing a better performance than with fixed PI and pole placement controllers

KW - adpative control

KW - case based reasoning

KW - Machine Learning

KW - Chemical reactors

KW - continuous stirred tank reactor

U2 - 10.1109/CONTROLO.2018.8514277

DO - 10.1109/CONTROLO.2018.8514277

M3 - Conference contribution

SN - 78-1-5386-5346-3

SP - 91

EP - 97

BT - 2018 13th APCA International Conference on Control and Soft Computing (CONTROLO)

ER -

Rocha T, Paredes S, Henriques J, Carvalho P, Cardoso A, Gil P. Adaptive Control Learning Based on a Similarity Measure. In 2018 13th APCA International Conference on Control and Soft Computing (CONTROLO). 2018. p. 91-97 https://doi.org/10.1109/CONTROLO.2018.8514277