Distributed intelligent illumination control in the context of probabilistic graphical models

M. Cosovic, T. Devaja, D. Bajovic, J. Machaj, G. McCutcheon, V. Stankovic, L. Stankovic, D. Vukobratovic

Research output: Contribution to conferencePaper

Abstract

Lighting systems based on light-emitting diodes (LEDs) possess many benefits over their incandescent counterparts including longer lifespans, lower energy costs, better quality of light and no toxic elements, all without sacrificing consumer satisfaction. Their lifespan is not affected by switching frequency allowing for better illumination control and system efficiency. In this paper, we present a fully distributed energy-saving illumination control strategy for the arrangements of a lighting network which consists of a group of LEDs and user-associated devices. All LEDs have a dimming feature in order to meet the illumination requirements of every user. In order to solve the optimization problem, we are using a distributed approach that utilizes factor graphs and the belief propagation algorithm. Using probabilistic graphical models to represent and solve the system model provides for a natural description of the problem structure, where user devices and LED controllers exchange data via line-of-sight communication.

Conference

Conference4th International Conference on Smart and Sustainable Technologies 2019
CountryCroatia
CitySplit
Period18/06/1921/06/19
Internet address

Fingerprint

Lighting
Light emitting diodes
Dimming (lamps)
Electronic data interchange
Switching frequency
Energy conservation
Controllers
Communication
Costs

Keywords

  • LED system
  • distributed control
  • newton method
  • factor graphs
  • belief propagation

Cite this

Cosovic, M., Devaja, T., Bajovic, D., Machaj, J., McCutcheon, G., Stankovic, V., ... Vukobratovic, D. (2019). Distributed intelligent illumination control in the context of probabilistic graphical models. Paper presented at 4th International Conference on Smart and Sustainable Technologies 2019 , Split, Croatia.
Cosovic, M. ; Devaja, T. ; Bajovic, D. ; Machaj, J. ; McCutcheon, G. ; Stankovic, V. ; Stankovic, L. ; Vukobratovic, D. / Distributed intelligent illumination control in the context of probabilistic graphical models. Paper presented at 4th International Conference on Smart and Sustainable Technologies 2019 , Split, Croatia.6 p.
@conference{c73c027c80d841d6878d961afff97e6f,
title = "Distributed intelligent illumination control in the context of probabilistic graphical models",
abstract = "Lighting systems based on light-emitting diodes (LEDs) possess many benefits over their incandescent counterparts including longer lifespans, lower energy costs, better quality of light and no toxic elements, all without sacrificing consumer satisfaction. Their lifespan is not affected by switching frequency allowing for better illumination control and system efficiency. In this paper, we present a fully distributed energy-saving illumination control strategy for the arrangements of a lighting network which consists of a group of LEDs and user-associated devices. All LEDs have a dimming feature in order to meet the illumination requirements of every user. In order to solve the optimization problem, we are using a distributed approach that utilizes factor graphs and the belief propagation algorithm. Using probabilistic graphical models to represent and solve the system model provides for a natural description of the problem structure, where user devices and LED controllers exchange data via line-of-sight communication.",
keywords = "LED system, distributed control, newton method, factor graphs, belief propagation",
author = "M. Cosovic and T. Devaja and D. Bajovic and J. Machaj and G. McCutcheon and V. Stankovic and L. Stankovic and D. Vukobratovic",
year = "2019",
month = "6",
day = "18",
language = "English",
note = "4th International Conference on Smart and Sustainable Technologies 2019 ; Conference date: 18-06-2019 Through 21-06-2019",
url = "http://2019.splitech.org/",

}

Cosovic, M, Devaja, T, Bajovic, D, Machaj, J, McCutcheon, G, Stankovic, V, Stankovic, L & Vukobratovic, D 2019, 'Distributed intelligent illumination control in the context of probabilistic graphical models' Paper presented at 4th International Conference on Smart and Sustainable Technologies 2019 , Split, Croatia, 18/06/19 - 21/06/19, .

Distributed intelligent illumination control in the context of probabilistic graphical models. / Cosovic, M.; Devaja, T.; Bajovic, D.; Machaj, J.; McCutcheon, G.; Stankovic, V.; Stankovic, L.; Vukobratovic, D.

2019. Paper presented at 4th International Conference on Smart and Sustainable Technologies 2019 , Split, Croatia.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Distributed intelligent illumination control in the context of probabilistic graphical models

AU - Cosovic, M.

AU - Devaja, T.

AU - Bajovic, D.

AU - Machaj, J.

AU - McCutcheon, G.

AU - Stankovic, V.

AU - Stankovic, L.

AU - Vukobratovic, D.

PY - 2019/6/18

Y1 - 2019/6/18

N2 - Lighting systems based on light-emitting diodes (LEDs) possess many benefits over their incandescent counterparts including longer lifespans, lower energy costs, better quality of light and no toxic elements, all without sacrificing consumer satisfaction. Their lifespan is not affected by switching frequency allowing for better illumination control and system efficiency. In this paper, we present a fully distributed energy-saving illumination control strategy for the arrangements of a lighting network which consists of a group of LEDs and user-associated devices. All LEDs have a dimming feature in order to meet the illumination requirements of every user. In order to solve the optimization problem, we are using a distributed approach that utilizes factor graphs and the belief propagation algorithm. Using probabilistic graphical models to represent and solve the system model provides for a natural description of the problem structure, where user devices and LED controllers exchange data via line-of-sight communication.

AB - Lighting systems based on light-emitting diodes (LEDs) possess many benefits over their incandescent counterparts including longer lifespans, lower energy costs, better quality of light and no toxic elements, all without sacrificing consumer satisfaction. Their lifespan is not affected by switching frequency allowing for better illumination control and system efficiency. In this paper, we present a fully distributed energy-saving illumination control strategy for the arrangements of a lighting network which consists of a group of LEDs and user-associated devices. All LEDs have a dimming feature in order to meet the illumination requirements of every user. In order to solve the optimization problem, we are using a distributed approach that utilizes factor graphs and the belief propagation algorithm. Using probabilistic graphical models to represent and solve the system model provides for a natural description of the problem structure, where user devices and LED controllers exchange data via line-of-sight communication.

KW - LED system

KW - distributed control

KW - newton method

KW - factor graphs

KW - belief propagation

UR - http://2019.splitech.org/

M3 - Paper

ER -

Cosovic M, Devaja T, Bajovic D, Machaj J, McCutcheon G, Stankovic V et al. Distributed intelligent illumination control in the context of probabilistic graphical models. 2019. Paper presented at 4th International Conference on Smart and Sustainable Technologies 2019 , Split, Croatia.