EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Seniors’ experiences in smart communities across day and night: Evidence from vision-language models and explainable AI

Yuzhuo Gao, Bo Xia, Chongxian Chen, Kirsty Volz, Mirko Guaralda, Jiaxuan E

Building and Environment · 2027-01 · 307 · 115302

DOI: 10.1016/j.buildenv.2026.115302 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Building and Environment, 2027-01. Verfasst von Yuzhuo Gao und 5 weiteren Autor:innen. Die Arbeit nennt 74 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: China Scholarship Council.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Age-friendly smart community design requires understanding how traditional and smart environmental features jointly shape older adults’ lived experience. However, existing research treats smart infrastructure as peripheral, prioritises discrete service functions over integrated experiential dimensions, and overlooks temporal variation between daytime and nighttime conditions. This study addresses these gaps by examining how environmental features influence seniors’ experience across seven dimensions, including accessibility, safety, health, stress reduction, sociability, belonging, and cultural identity, under contrasting temporal contexts. Focusing on Yuexiu District, Guangzhou, we analysed 56,479 daytime and 55,310 nighttime street-view images, extracting features via semantic segmentation, HSV-based indicators, and vision-language models detecting 20 smart and design feature categories. Experience scores were derived from pairwise evaluations by 152 older adults and spatially predicted using a deep learning ResNet50 model. Nonlinear relationships were quantified using XGBoost-SHAP-GAM. Four key findings emerge: (a) experience outcomes exhibit dimension-specific spatial configurations, with relatively coherent daytime patterns becoming more fragmented at night; (b) dominant environmental mechanisms shift from spatial composition during the day to illumination and visual richness at night; (c) environmental effects operate within narrow nonlinear sensitivity boundaries, where small changes can reverse their direction; and (d) smart features produce type-specific and temporally contingent effects rather than uniformly positive outcomes, with some interventions (e.g., surveillance cameras) reducing rather than enhancing experiential quality. These findings provide empirical evidence for inclusive smart ageing communities.

Förderung

  • China Scholarship Council

Zitieren

APA

Gao, Y., Xia, B., Chen, C., Volz, K., Guaralda, M., & E, J. (2027). Seniors’ experiences in smart communities across day and night: Evidence from vision-language models and explainable AI. Building and Environment, 307, 115302. https://doi.org/10.1016/j.buildenv.2026.115302

Vancouver

Gao Y, Xia B, Chen C, Volz K, Guaralda M, E J. Seniors’ experiences in smart communities across day and night: Evidence from vision-language models and explainable AI. Building and Environment. 2027;307:115302. doi:10.1016/j.buildenv.2026.115302

BibTeX
@article{Gao2027,
  author = {Gao, Yuzhuo and Xia, Bo and Chen, Chongxian and Volz, Kirsty and Guaralda, Mirko and E, Jiaxuan},
  title = {Seniors’ experiences in smart communities across day and night: Evidence from vision-language models and explainable AI},
  journal = {Building and Environment},
  year = {2027},
  volume = {307},
  pages = {115302},
  doi = {10.1016/j.buildenv.2026.115302}
}
Kostenlos starten →