Publication

Abstract : Crowdsourcing is a mechanism of outsourcing a task to a multitude of people. It includes workers, tasks and a web-based platform to intermediate between the tasks and workers. Spatial crowdsourcing is a paradigm of crowdsourcing, that extends beyond the traditional web-based crowdsourcing by including physical entities of the geographic system. This data could be in the form of topology, location or longitude, and latitude coordinates. The task allocation on a spatial crowdsourcing platform based on proximity between task and worker is challenging since the worker must be physically present at the location to complete the task. Existing methods use Euclidean distance measure for task allocation. However, it offers less accuracy when applied in a 3D plane. Hence, this paper proposes a novel task allocation scheme through a score calculation that uses Haversine distance measure incorporated with worker preference and reputation for allocating the task. We validate the proposed task allocation model by comparing the task allocation cost between 2D and 3D planes, using real and synthetic data. Our observations clearly display considerable cost-benefit and increased task allocation in our proposed solution.