Reinforcement Learning-based Thermal Comfort Control for Vehicle Cabins

Research output: Contribution to journalArticle

1 Citation (Scopus)
2 Downloads (Pure)

Abstract

Vehicle climate control systems aim to keep passengers thermally comfortable. However, current systems control temperature rather than thermal comfort and tend to be energy hungry, which is of particular concern when considering electric vehicles. This paper poses energy-efficient vehicle comfort control as a Markov Decision Process, which is then solved numerically using Sarsa(λ) and an empirically validated, single-zone, 1D thermal model of the cabin. The resulting controller was tested in simulation using 200 randomly selected scenarios and found to exceed the performance of bang-bang, proportional, simple fuzzy logic, and commercial controllers with 23%, 43%, 40%, 56% increase, respectively. Compared to the next best performing controller, energy consumption is reduced by 13% while the proportion of time spent thermally comfortable is increased by 23%. These results indicate that this is a viable approach that promises to translate into substantial comfort and energy improvements in the car.
Original languageEnglish
Pages (from-to)413-421
JournalMechatronics
Volume50
Early online date3 May 2017
DOIs
Publication statusPublished - Apr 2018

Fingerprint

Thermal comfort
Reinforcement learning
Controllers
Climate control
Control systems
Electric vehicles
Temperature control
Fuzzy logic
Railroad cars
Energy utilization

Keywords

  • Thermal comfort
  • Reinforcement learning
  • Equivalent temperature
  • Comfort model
  • Energy consumption

Cite this

Reinforcement Learning-based Thermal Comfort Control for Vehicle Cabins. / Brusey, James; Hintea, Diana; Gaura, Elena; Beloe, Neil.

In: Mechatronics, Vol. 50, 04.2018, p. 413-421.

Research output: Contribution to journalArticle

@article{d87b95b2f7c44d378a0cffefc31c3507,
title = "Reinforcement Learning-based Thermal Comfort Control for Vehicle Cabins",
abstract = "Vehicle climate control systems aim to keep passengers thermally comfortable. However, current systems control temperature rather than thermal comfort and tend to be energy hungry, which is of particular concern when considering electric vehicles. This paper poses energy-efficient vehicle comfort control as a Markov Decision Process, which is then solved numerically using Sarsa(λ) and an empirically validated, single-zone, 1D thermal model of the cabin. The resulting controller was tested in simulation using 200 randomly selected scenarios and found to exceed the performance of bang-bang, proportional, simple fuzzy logic, and commercial controllers with 23{\%}, 43{\%}, 40{\%}, 56{\%} increase, respectively. Compared to the next best performing controller, energy consumption is reduced by 13{\%} while the proportion of time spent thermally comfortable is increased by 23{\%}. These results indicate that this is a viable approach that promises to translate into substantial comfort and energy improvements in the car.",
keywords = "Thermal comfort, Reinforcement learning, Equivalent temperature, Comfort model, Energy consumption",
author = "James Brusey and Diana Hintea and Elena Gaura and Neil Beloe",
year = "2018",
month = "4",
doi = "10.1016/j.mechatronics.2017.04.010",
language = "English",
volume = "50",
pages = "413--421",
journal = "Mechatronics",
issn = "0957-4158",
publisher = "Elsevier",

}

TY - JOUR

T1 - Reinforcement Learning-based Thermal Comfort Control for Vehicle Cabins

AU - Brusey, James

AU - Hintea, Diana

AU - Gaura, Elena

AU - Beloe, Neil

PY - 2018/4

Y1 - 2018/4

N2 - Vehicle climate control systems aim to keep passengers thermally comfortable. However, current systems control temperature rather than thermal comfort and tend to be energy hungry, which is of particular concern when considering electric vehicles. This paper poses energy-efficient vehicle comfort control as a Markov Decision Process, which is then solved numerically using Sarsa(λ) and an empirically validated, single-zone, 1D thermal model of the cabin. The resulting controller was tested in simulation using 200 randomly selected scenarios and found to exceed the performance of bang-bang, proportional, simple fuzzy logic, and commercial controllers with 23%, 43%, 40%, 56% increase, respectively. Compared to the next best performing controller, energy consumption is reduced by 13% while the proportion of time spent thermally comfortable is increased by 23%. These results indicate that this is a viable approach that promises to translate into substantial comfort and energy improvements in the car.

AB - Vehicle climate control systems aim to keep passengers thermally comfortable. However, current systems control temperature rather than thermal comfort and tend to be energy hungry, which is of particular concern when considering electric vehicles. This paper poses energy-efficient vehicle comfort control as a Markov Decision Process, which is then solved numerically using Sarsa(λ) and an empirically validated, single-zone, 1D thermal model of the cabin. The resulting controller was tested in simulation using 200 randomly selected scenarios and found to exceed the performance of bang-bang, proportional, simple fuzzy logic, and commercial controllers with 23%, 43%, 40%, 56% increase, respectively. Compared to the next best performing controller, energy consumption is reduced by 13% while the proportion of time spent thermally comfortable is increased by 23%. These results indicate that this is a viable approach that promises to translate into substantial comfort and energy improvements in the car.

KW - Thermal comfort

KW - Reinforcement learning

KW - Equivalent temperature

KW - Comfort model

KW - Energy consumption

U2 - 10.1016/j.mechatronics.2017.04.010

DO - 10.1016/j.mechatronics.2017.04.010

M3 - Article

VL - 50

SP - 413

EP - 421

JO - Mechatronics

JF - Mechatronics

SN - 0957-4158

ER -