From sensing to controlling: The state of the art in ubiquitous crowdsourcing

Crowdsourcing systems aim to solve complex tasks by engaging a multitude of humans through mobile and social media and by harnessing their ability to perform specific subtasks and to report the results. Given the rapid increase in the number, size and diversity of ubiquitous crowdsourcing applications across a wide spectrum of domains all around the world, there is a need for a principled way of conceptualising such systems. The first step is the development of a crowdsourcing typology to create a baseline for comparing, evaluating and advancing efforts in this field. For example, crowdsourcing systems need to address a common set of issues to ensure their effectiveness, such as quality control and incentive design. In this paper, we present a thematic overview of the emerging ubiquitous crowdsourcing systems by grouping them based on the task complexity that they support and along key challenges in successfully harvesting expertise of large human networks.