A Human Sensing Framework for Smart Buildings Facilities

Abstract

Faults and incidents can occur in building facilities and detecting and reporting them is considered a challenge. Building managers need to have real-time condition updates about the facility or potential incidents (e.g., loose door handles, sensors not working affecting lights and temperature) that might lead to safety and security losses if they are not detected and fixed. As buildings become smarter, technological solutions can help, but can equally be the source of faults and incidents. Human participation is still needed to help have adequate awareness of the situation in the building environment, but building occupants may not always spontaneously report faults as quickly and reliably as needed. Crowdsourcing is the practice of obtaining information by recruiting people’s knowledge. Previous studies have shown that crowdsourced contributions, which are created by nonexpert observations and interpretations, are fairly similar to the experts’ outputs, especially for nonsensitive data, and therefore be an alternative to hiring experts or dedicated inspectors. Crowdsourcing, however, requires suitable motivational mechanisms and an appropriate design, and its applicability as a viable solution to detect and report faults in smart buildings has not been studied. To study that, I used a form of crowdsourcing known as human sensing which includes individuals contributing their observations within their local area like their neighbourhoods, workplaces, and often visited places. This thesis provides three contributions to the application of crowdsourcing for smart buildings. First, perceptions and motivations that drive occupants to engage with smart building crowdsourcing, are informed by conducting semi-structured interviews. Second, human sensing feasibility is obtained by running a real-world experiment. Third, a framework describing a set of guidelines covering the main dimensions known in the crowdsourcing process: platform, requesters, workers, incentives, task, and environmental preparation. It was done by synthesising the results and observations discovered from the two previous studies and related research. These contributions enabled me to conclude that crowdsourcing can be a good option for reporting faults in smart buildings if a set of guidelines are followed to deploy the initiative successfully.