Show simple item record

dc.contributor.authorMontanari, Alessandro
dc.date.accessioned2019-05-08T11:15:07Z
dc.date.available2019-05-08T11:15:07Z
dc.date.submitted2018-05-10
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/292527
dc.description.abstractThe importance of studying social interactions has been proven useful in several fields. In the workplace, studies have found that allowing mixing among different groups could improve team coordination and productivity. Architectural studies have analysed how physical spaces can potentially increase unplanned interactions. Other areas such as epidemiology have also benefited from tracking face-to-face contacts to study the spread of disease. Although technology has progressed significantly, the automated and accurate measurement of human interactions with mobile devices is still lagging. The main shortcomings have to do with accuracy of the captured data and with the communication modalities considered. Additionally, non-verbal behaviours during social interactions (e.g. body posture, orientation and interaction distance) have been often neglected, with a few exceptions, even if traditional sociology has highlighted their importance. In this dissertation we address these challenges by developing two wearable research platforms to monitor different dimensions of social interactions. First, we study the extent to which Bluetooth Low Energy could detect proximity in indoor environments. We analyse all the relevant protocol parameters and measure their impact on power consumption, on custom as well as on commercial devices. We assess its accuracy with a 4-week long deployment illustrating its sustainability for social dynamics studies. With the contacts and mobility data collected during the deployment we study the relationship between social contacts and space design, focusing on a modern architectural concept, Activity-Based Working (ABW). We uncover several patterns and we show how they could be the result of the correct adoption of ABW principles. However, we also discover that the employees might not have fully embraced the ABW concepts entirely, leading to mismatches between principles and actual space usage. Given the importance of studying non-verbal behaviour during social contact we then devise a novel wearable device that, by exploiting near-infrared signals, is able to capture accurate information about distance and angle of interaction between people. We show how we design the device to be robust to ambient light changes and short occlusions by leveraging inertial measurement units. With extensive testing we evaluate its accuracy and robustness. We then explore its potential to study creative processes by deploying it to capture non-verbal cues during a creative task. We show how data about the relative orientation between people and their interpersonal distance could be used to predict the role they have during the interaction and the status of the task. The platforms developed and the insights drawn in this dissertation provide evidence to support the use of wearable technologies to monitor social interactions at an unprecedented level.
dc.language.isoen
dc.rightsAll rights reserved
dc.subjectwearables
dc.subjectsocial dynamics
dc.subjectnon-verbal behaviour monitoring
dc.subjectsocial dynamics monitoring
dc.subjectcomputational social science
dc.subjectsocial interactions
dc.titleDevising and evaluating wearable technology for social dynamics monitoring
dc.typeThesis
dc.type.qualificationlevelDoctoral
dc.type.qualificationnameDoctor of Philosophy (PhD)
dc.publisher.institutionUniversity of Cambridge
dc.publisher.departmentDepartment of Computer Science and Technology
dc.date.updated2019-05-08T10:25:47Z
dc.identifier.doi10.17863/CAM.39687
dc.publisher.collegeGirton College
dc.type.qualificationtitlePhD in Computer Science
cam.supervisorMascolo, Cecilia
cam.thesis.fundingtrue


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record