Toward Privacy-Preserving Task Assignment for Fully Distributed Spatial Crowdsourcing

With the proliferation of human-carried mobile devices, spatial crowdsourcing has emerged as a transformative system, where requesters outsource their spatiotemporal tasks to a set of workers who are willing to perform the tasks at the specified locations. However, in order to make efficient assignments, the existing spatial crowdsourcing system usually requires workers and/or tasks to expose their locations, which raises a significant concern of compromising location privacy. In addition, traditional spatial crowdsourcing systems employ a centralized server to manage the information of workers and tasks. Such a centralized design does not scale to a large number of workers/tasks, making the server easily a bottleneck. In this article, we present an online framework for assigning tasks to workers without compromising the location privacy in a fully distributed manner. Our system protects the location privacy of both workers and tasks through homomorphic encryption. We further propose a novel wait-and-decide mechanism and a proportional-backoff mechanism to increase the number of assigned tasks. Extensive experiments on real-world data sets illustrate that our proposed system achieves a large number of task assignments in an efficient and privacy-preserving manner.