Emerging spatial crowdsourcing (SC) provides an approach for collecting and analyzing spatiotemporal information from intelligent transportation systems. However, the exposure of massive location privacy to potential adversaries for the purpose of quality control makes workers more vulnerable. To protect workers’ location privacy, an obfuscation scheme is proposed to incorporate uncertainties into the SC quality control problem through obfuscating the standard location data in terms of both space and time. Two measures, location entropy and results accuracy, are used to evaluate the performance of location privacy protection. We theoretically and experimentally confirm the security and accuracy of the obfuscation approach. The results of experiments show that: a) hiding workers’ location from the requester reduces the quality of SC; and b) obfuscation arithmetic with appropriate obfuscation coefficients protects workers’ location privacy with little effect on SC quality. Under the protection of this obfuscation scheme, the new system provides better security and similar quality compared to the existing SC system.

Introduction

The crowdsourcing model is frequently used to gather data in intelligent transportation systems (ITS) applications e.g., avoiding traffic congestion. The new mechanism for collecting and analyzing spatiotemporal information is spatial crowdsourcing (SC), and this mechanism exploits a large volume of vehicles and their mobility. In the SC platform, a requester outsources a set of spatiotemporal tasks to a set of workers with mobile devices, and workers perform tasks after physically traveling to places of interest [1], e.g. the outsourcer requests workers in cars to collect real-time traffic information on roads Moreover, to assist in checking workers’ submissions, some existing SC platforms require workers to disclose their immediate locations along with the taskspecified submission to the requester, who may be a potential adversary seeking to attack the location privacy of individual workers [2]. In practice, some malicious requesters may collect private information on worker locations through deliberately designed SC tasks. In addition, a crowdsourcing task is a kind of micro task, so a worker may submit numerous tasks with location information in a short period of time. Revealing workers’ precise locations may allow an adversary to infer sensitive information and even to stalk or mug workers [3]. Hence, protecting location privacy is an essential aspect of SC, since workers will not agree to participate in spatial tasks if there is a possibility of a privacy breach.