Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference

Mohamed Baioumy; Corrado Pezzato; Carlos Hernández Corbato; Nick Hawes; Riccardo Ferrari

doi:10.1007/978-3-030-93736-2_48

Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference

Mohamed Baioumy^*, Corrado Pezzato, Carlos Hernández Corbato, Nick Hawes, Riccardo Ferrari

^*Corresponding author for this work

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

3 Citations (Scopus)

57 Downloads (Pure)

Abstract

This work presents a fault-tolerant control scheme for sensory faults in robotic manipulators based on active inference. In the majority of existing schemes a binary decision of whether a sensor is healthy (functional) or faulty is made based on measured data. The decision boundary is called a threshold and it is usually deterministic. Following a faulty decision, fault recovery is obtained by excluding the malfunctioning sensor. We propose a stochastic fault-tolerant scheme based on active inference and precision learning which does not require a priori threshold definitions to trigger fault recovery. Instead, the sensor precision, which represents its health status, is learned online in a model-free way allowing the system to gradually, and not abruptly exclude a failing unit. Experiments on a robotic manipulator show promising results and directions for future work are discussed.

Original language	English
Title of host publication	Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Subtitle of host publication	Proceedings of the International Workshops of ECML PKDD 2021
Editors	Michael Kamp, Michael Kamp, Irena Koprinska, et. al.
Publisher	Springer
Pages	681-691
ISBN (Print)	978-3-030-93735-5
DOIs	https://doi.org/10.1007/978-3-030-93736-2_48
Publication status	Published - 2022
Event	21st European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021 - Virtual, Online Duration: 13 Sept 2021 → 17 Sept 2021

Publication series

Name	Communications in Computer and Information Science
Volume	1524 CCIS
ISSN (Print)	1865-0929
ISSN (Electronic)	1865-0937

Conference

Conference	21st European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021
City	Virtual, Online
Period	13/09/21 → 17/09/21

Bibliographical note

Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Access to Document

10.1007/978-3-030-93736-2_48

Baioumy2021_Chapter_TowardsStochasticFault-ToleranFinal published version, 415 KB

Cite this

Baioumy, M., Pezzato, C., Corbato, C. H., Hawes, N., & Ferrari, R. (2022). Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference. In M. Kamp, M. Kamp, I. Koprinska, & E. A. (Eds.), Machine Learning and Principles and Practice of Knowledge Discovery in Databases: Proceedings of the International Workshops of ECML PKDD 2021 (pp. 681-691). (Communications in Computer and Information Science; Vol. 1524 CCIS). Springer. https://doi.org/10.1007/978-3-030-93736-2_48

Baioumy, Mohamed ; Pezzato, Corrado ; Corbato, Carlos Hernández et al. / Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference. Machine Learning and Principles and Practice of Knowledge Discovery in Databases: Proceedings of the International Workshops of ECML PKDD 2021. editor / Michael Kamp ; Michael Kamp ; Irena Koprinska ; et. al. Springer, 2022. pp. 681-691 (Communications in Computer and Information Science).

@inproceedings{d3e3557ed79e4e6f93f8965c514af55e,

title = "Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference",

abstract = "This work presents a fault-tolerant control scheme for sensory faults in robotic manipulators based on active inference. In the majority of existing schemes a binary decision of whether a sensor is healthy (functional) or faulty is made based on measured data. The decision boundary is called a threshold and it is usually deterministic. Following a faulty decision, fault recovery is obtained by excluding the malfunctioning sensor. We propose a stochastic fault-tolerant scheme based on active inference and precision learning which does not require a priori threshold definitions to trigger fault recovery. Instead, the sensor precision, which represents its health status, is learned online in a model-free way allowing the system to gradually, and not abruptly exclude a failing unit. Experiments on a robotic manipulator show promising results and directions for future work are discussed.",

author = "Mohamed Baioumy and Corrado Pezzato and Corbato, {Carlos Hern{\'a}ndez} and Nick Hawes and Riccardo Ferrari",

note = "Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.; 21st European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021 ; Conference date: 13-09-2021 Through 17-09-2021",

year = "2022",

doi = "10.1007/978-3-030-93736-2_48",

language = "English",

isbn = "978-3-030-93735-5",

series = "Communications in Computer and Information Science",

publisher = "Springer",

pages = "681--691",

editor = "Michael Kamp and Michael Kamp and Irena Koprinska and {et. al.}",

booktitle = "Machine Learning and Principles and Practice of Knowledge Discovery in Databases",

}

Baioumy, M, Pezzato, C, Corbato, CH, Hawes, N & Ferrari, R 2022, Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference. in M Kamp, M Kamp, I Koprinska & EA (eds), Machine Learning and Principles and Practice of Knowledge Discovery in Databases: Proceedings of the International Workshops of ECML PKDD 2021. Communications in Computer and Information Science, vol. 1524 CCIS, Springer, pp. 681-691, 21st European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021, Virtual, Online, 13/09/21. https://doi.org/10.1007/978-3-030-93736-2_48

Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference. / Baioumy, Mohamed; Pezzato, Corrado; Corbato, Carlos Hernández et al.
Machine Learning and Principles and Practice of Knowledge Discovery in Databases: Proceedings of the International Workshops of ECML PKDD 2021. ed. / Michael Kamp; Michael Kamp; Irena Koprinska; et. al. Springer, 2022. p. 681-691 (Communications in Computer and Information Science; Vol. 1524 CCIS).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference

AU - Baioumy, Mohamed

AU - Pezzato, Corrado

AU - Corbato, Carlos Hernández

AU - Hawes, Nick

AU - Ferrari, Riccardo

N1 - Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2022

Y1 - 2022

N2 - This work presents a fault-tolerant control scheme for sensory faults in robotic manipulators based on active inference. In the majority of existing schemes a binary decision of whether a sensor is healthy (functional) or faulty is made based on measured data. The decision boundary is called a threshold and it is usually deterministic. Following a faulty decision, fault recovery is obtained by excluding the malfunctioning sensor. We propose a stochastic fault-tolerant scheme based on active inference and precision learning which does not require a priori threshold definitions to trigger fault recovery. Instead, the sensor precision, which represents its health status, is learned online in a model-free way allowing the system to gradually, and not abruptly exclude a failing unit. Experiments on a robotic manipulator show promising results and directions for future work are discussed.

AB - This work presents a fault-tolerant control scheme for sensory faults in robotic manipulators based on active inference. In the majority of existing schemes a binary decision of whether a sensor is healthy (functional) or faulty is made based on measured data. The decision boundary is called a threshold and it is usually deterministic. Following a faulty decision, fault recovery is obtained by excluding the malfunctioning sensor. We propose a stochastic fault-tolerant scheme based on active inference and precision learning which does not require a priori threshold definitions to trigger fault recovery. Instead, the sensor precision, which represents its health status, is learned online in a model-free way allowing the system to gradually, and not abruptly exclude a failing unit. Experiments on a robotic manipulator show promising results and directions for future work are discussed.

UR - http://www.scopus.com/inward/record.url?scp=85121019036&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-93736-2_48

DO - 10.1007/978-3-030-93736-2_48

M3 - Conference contribution

AN - SCOPUS:85121019036

SN - 978-3-030-93735-5

T3 - Communications in Computer and Information Science

SP - 681

EP - 691

BT - Machine Learning and Principles and Practice of Knowledge Discovery in Databases

A2 - Kamp, Michael

A2 - Koprinska, Irena

A2 - null, et. al.

PB - Springer

T2 - 21st European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021

Y2 - 13 September 2021 through 17 September 2021

ER -

Baioumy M, Pezzato C, Corbato CH, Hawes N, Ferrari R. Towards Stochastic Fault-Tolerant Control Using Precision Learning and Active Inference. In Kamp M, Kamp M, Koprinska I, EA, editors, Machine Learning and Principles and Practice of Knowledge Discovery in Databases: Proceedings of the International Workshops of ECML PKDD 2021. Springer. 2022. p. 681-691. (Communications in Computer and Information Science). doi: 10.1007/978-3-030-93736-2_48