打印

作者

Francisco Benita;Bige Tunçer

来源

URBAN FORESTRY & URBAN GREENING,Vol.43,Issue1,Article 126365

语言

英文

关键字

Urban spaces;Immediate environment;Stress;Green spaces;Multi-sensor data fusion

作者单位

Architecture and Sustainable Design, Singapore University of Technology and Design, 8 Somapah Road, Singapore 487372, Singapore;Architecture and Sustainable Design, Singapore University of Technology and Design, 8 Somapah Road, Singapore 487372, Singapore

摘要

This study investigates the relationship between urban features (sky exposure, green spaces, visual complexity, and built-up area), immediate environmental factors (air temperature, relative humidity, Heat Stress Index, Wet Bulb Globe Temperature, wind speed, and noise), personal characteristics (perceived restorativeness) and body reactions (body skin temperature and skin conductance responses). The proposed framework is based on multi-sensor data fusion from wearable physiological sensors, wireless environmental sensors, smartphones, images, geographic information systems datasets, and questionnaires. An experimental setup in a real-world setting is conducted and machine learning algorithms for regression problems and feature selection for variable importance are implemented. The results suggest a significant association between immediate environmental factors and body reactions; however, urban features are found to be weak explanatory variables. A deeper analysis of the identified stress hotspots revealed that locations with more dense green spaces, greater sky exposure, and smaller built-up area tended to report lower levels of stress reaction.