Location privacy-aware task recommendation for spatial crowdsourcing

Abstract

Spatial crowdsourcing engages individuals to collect and process social, environmental and other information with spatio-temporal features, making the data collection and analysis efficient, scalable and smart. The quality of task fulfillment strongly depends on the set of recruited workers. The more suitable workers are engaged, the better results may be obtained, meanwhile, the more privacy of workers will be disclosed. In this paper, we propose LATE, a novel location privacy-aware task recommendation framework in spatial crowdsourcing, which enables spatial crowdsourcing servers (SC-servers) to recommend spatial tasks released by customers to the workers in geocast regions. Based on Lagrange Interpolating Polynomials, we design a privacy-preserving location matching mechanism to allow the SC-server to determine whether a worker is in geocast region of a spatial task or not without any knowledge about the task's geocast region and the worker's location. In addition, the spatial tasks and crowdsourcing reports are protected against privacy leakage for both customers and workers. Finally, we discuss the security properties of LATE and demonstrate its efficiency on computation and communication.