Please use this identifier to cite or link to this item:
Type: Conference paper
Title: A deep learning approach to personal thermal comfort models for an ageing population
Author: Arakawa Martins, L.
Soebarto, V.
Williamson, T.
Pisaniello, D.
Citation: Proceedings of the 54th International Conference of the Architectural Science Association (ANZAScA) 2020. Imaginable Futures: Design Thinking, and the Scientific Method, 2020 / Ghaffarianhoseini, A. (ed./s), vol.2020-November, pp.71-80
Publisher: Architectural Science Association (ANZAScA)
Publisher Place: online
Issue Date: 2020
ISBN: 9780992383572
ISSN: 2209-3850
Conference Name: International Conference of the Architectural Science Association (ANZAScA) (26 Nov 2020 - 27 Nov 2020 : virtual online)
Editor: Ghaffarianhoseini, A.
Statement of
Larissa Arakawa Martins, Veronica Soebarto, Terence Williamson and Dino Pisaniello
Abstract: Recent years have shown an increasing number of studies on personal thermal comfort models as an alternative to the conventional approach to understanding thermal comfort in the built environment. Instead of basing on an average response from a large population, personalized models are designed to predict individuals’ thermal comfort responses, using a person’s direct feedback and personal characteristics as calibration inputs. However, personal comfort models have mainly used data from office environments and healthy younger adults. Studies on personal comfort models that focus on older people and dwellings are still absent in the literature. Nonetheless, considering the worldwide changing climate, the ageing population and older people’s heterogeneity in terms of intrinsic capacities and needs, personalized models could be the most appropriate path towards recognizing diversity and predicting individual thermal preferences in a more accurate way. This paper shows examples of personal comfort models, using deep learning algorithms and environmental and personal characteristics as inputs, derived from an on-going study that monitored people aged 65 and over in South Australia who live at home. The results have so far indicated that, on average, the individualised models improved the predictions by 69% when compared to traditional models.
Keywords: Personal comfort models
machine learning
thermal comfort
older people
Rights: © 2020 and published by the Architectural Science Association (ANZAScA).
Grant ID:
Published version:
Appears in Collections:Architecture publications

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.