Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications

Michael A. Demetriou, Antonios Armaou

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

2 Citations (Scopus)

Abstract

We consider the problem of sensor placement (both spatial location and number) for a class of parameter-dependent diffusion-advection processes that model environmental processes. To distinguish between advection and diffusion dominated environmental processes, the Péclet number, Pe, which takes values over a set of a priori defined values, is utilized to optimize the number and spatial location of the sensors required. The minimum number of sensors is defined using system theoretic measures and essentially considers the smallest number of sensors that would render the system observable, thereby facilitating the design of a state observer. The optimization metric is defined with respect to the spatial H2 norm of the dominant system modes, which may differ for different values of Pe. For each value of Pe, a set of optimal sensor locations and number is found and the associated state estimator is designed. The supervisory scheme then schedules the sensors corresponding to the Péclet number that describes the process at a given time by pouting in sleep mode all sensors associated with a different value of Pe and activating the sensors that are optimal for the current value of Pe. At the same time, the state estimator also switches by using the filter gain corresponding to the current value of the Péclet number and the active sensors. Extensive simulation studies are included to provide further inside on parameter-dependent sensor and observer scheduling for environmental processes.

Original languageEnglish (US)
Title of host publication2014 American Control Conference, ACC 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages4056-4062
Number of pages7
ISBN (Print)9781479932726
DOIs
StatePublished - Jan 1 2014
Event2014 American Control Conference, ACC 2014 - Portland, OR, United States
Duration: Jun 4 2014Jun 6 2014

Other

Other2014 American Control Conference, ACC 2014
CountryUnited States
CityPortland, OR
Period6/4/146/6/14

Fingerprint

Scheduling
Sensors
Advection
Switches

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering

Cite this

Demetriou, M. A., & Armaou, A. (2014). Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications. In 2014 American Control Conference, ACC 2014 (pp. 4056-4062). [6859457] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ACC.2014.6859457
Demetriou, Michael A. ; Armaou, Antonios. / Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications. 2014 American Control Conference, ACC 2014. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 4056-4062
@inproceedings{32694579c6d34033999fc7c4c14bd4b1,
title = "Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications",
abstract = "We consider the problem of sensor placement (both spatial location and number) for a class of parameter-dependent diffusion-advection processes that model environmental processes. To distinguish between advection and diffusion dominated environmental processes, the P{\'e}clet number, Pe, which takes values over a set of a priori defined values, is utilized to optimize the number and spatial location of the sensors required. The minimum number of sensors is defined using system theoretic measures and essentially considers the smallest number of sensors that would render the system observable, thereby facilitating the design of a state observer. The optimization metric is defined with respect to the spatial H2 norm of the dominant system modes, which may differ for different values of Pe. For each value of Pe, a set of optimal sensor locations and number is found and the associated state estimator is designed. The supervisory scheme then schedules the sensors corresponding to the P{\'e}clet number that describes the process at a given time by pouting in sleep mode all sensors associated with a different value of Pe and activating the sensors that are optimal for the current value of Pe. At the same time, the state estimator also switches by using the filter gain corresponding to the current value of the P{\'e}clet number and the active sensors. Extensive simulation studies are included to provide further inside on parameter-dependent sensor and observer scheduling for environmental processes.",
author = "Demetriou, {Michael A.} and Antonios Armaou",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/ACC.2014.6859457",
language = "English (US)",
isbn = "9781479932726",
pages = "4056--4062",
booktitle = "2014 American Control Conference, ACC 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

Demetriou, MA & Armaou, A 2014, Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications. in 2014 American Control Conference, ACC 2014., 6859457, Institute of Electrical and Electronics Engineers Inc., pp. 4056-4062, 2014 American Control Conference, ACC 2014, Portland, OR, United States, 6/4/14. https://doi.org/10.1109/ACC.2014.6859457

Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications. / Demetriou, Michael A.; Armaou, Antonios.

2014 American Control Conference, ACC 2014. Institute of Electrical and Electronics Engineers Inc., 2014. p. 4056-4062 6859457.

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

TY - GEN

T1 - Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications

AU - Demetriou, Michael A.

AU - Armaou, Antonios

PY - 2014/1/1

Y1 - 2014/1/1

N2 - We consider the problem of sensor placement (both spatial location and number) for a class of parameter-dependent diffusion-advection processes that model environmental processes. To distinguish between advection and diffusion dominated environmental processes, the Péclet number, Pe, which takes values over a set of a priori defined values, is utilized to optimize the number and spatial location of the sensors required. The minimum number of sensors is defined using system theoretic measures and essentially considers the smallest number of sensors that would render the system observable, thereby facilitating the design of a state observer. The optimization metric is defined with respect to the spatial H2 norm of the dominant system modes, which may differ for different values of Pe. For each value of Pe, a set of optimal sensor locations and number is found and the associated state estimator is designed. The supervisory scheme then schedules the sensors corresponding to the Péclet number that describes the process at a given time by pouting in sleep mode all sensors associated with a different value of Pe and activating the sensors that are optimal for the current value of Pe. At the same time, the state estimator also switches by using the filter gain corresponding to the current value of the Péclet number and the active sensors. Extensive simulation studies are included to provide further inside on parameter-dependent sensor and observer scheduling for environmental processes.

AB - We consider the problem of sensor placement (both spatial location and number) for a class of parameter-dependent diffusion-advection processes that model environmental processes. To distinguish between advection and diffusion dominated environmental processes, the Péclet number, Pe, which takes values over a set of a priori defined values, is utilized to optimize the number and spatial location of the sensors required. The minimum number of sensors is defined using system theoretic measures and essentially considers the smallest number of sensors that would render the system observable, thereby facilitating the design of a state observer. The optimization metric is defined with respect to the spatial H2 norm of the dominant system modes, which may differ for different values of Pe. For each value of Pe, a set of optimal sensor locations and number is found and the associated state estimator is designed. The supervisory scheme then schedules the sensors corresponding to the Péclet number that describes the process at a given time by pouting in sleep mode all sensors associated with a different value of Pe and activating the sensors that are optimal for the current value of Pe. At the same time, the state estimator also switches by using the filter gain corresponding to the current value of the Péclet number and the active sensors. Extensive simulation studies are included to provide further inside on parameter-dependent sensor and observer scheduling for environmental processes.

UR - http://www.scopus.com/inward/record.url?scp=84905714724&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84905714724&partnerID=8YFLogxK

U2 - 10.1109/ACC.2014.6859457

DO - 10.1109/ACC.2014.6859457

M3 - Conference contribution

AN - SCOPUS:84905714724

SN - 9781479932726

SP - 4056

EP - 4062

BT - 2014 American Control Conference, ACC 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Demetriou MA, Armaou A. Frequency domain methods for optimal sensor placement and scheduling of spatially distributed systems arising in environmental and meteorological applications. In 2014 American Control Conference, ACC 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 4056-4062. 6859457 https://doi.org/10.1109/ACC.2014.6859457