Aarhus Universitets segl

Pilot study: the feasibility of integrating multiple personalised sensors and machine learning for measuring urban features and body responses

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review

New emerging wearable sensors provide an opportunity to assist in the objective monitoring of human’s exposure in the urban environment. However, very few studies have applied wearable cameras together with health trackers to assess the effect of urban features on individuals in the built environment. This paper proposes a new approach and applies it in a pilot study to test its feasibility. This approach employes a FrontRow (FR) wearable lifestyle camera, a GPS tracker and an Empatica 4 (E4) wristband as a sensor package to track individuals during their real-life activities. Then, machine learning methods are adopted to analyse the imagery from wearable cameras to extract urban features, using Microsoft API of cognitive service and SegNet model to detect urban features, and then linking up these features with physiological data (e.g., electrodermal activity) from E4 and location data from a GPS device to indicate the correlation of spatial distribution. In this pilot study, volunteers (k=12) were recruited in November and December 2020 and asked to conduct a self-led city tour (average length = 2h, 30min) with the sensors around the centre of Roskilde, Denmark. Simultaneously, two of the volenteers wore the sensors for one week. At the end of the pilot campaign, the results prove the feasibility of our proposed approach and show the potentials of integrated, multi-sourced data in the study of urban health.
TidsskriftInternational Journal of E-Planning Research
StatusAfsendt - 2022

Se relationer på Aarhus Universitet Citationsformater

ID: 276236161