Location privacy protection method based on differential privacy in crowdsensing task allocation
Qiong Zhang,Taochun Wang,Yuan Tao,Nuo Xu,Fulong Chen,Dong Xie
DOI: https://doi.org/10.1016/j.adhoc.2024.103464
IF: 4.816
2024-03-06
Ad Hoc Networks
Abstract:With the widespread popularity of smart phones, watches and other devices, mobile crowdsensing (MCS) has gradually entered the public's field of vision. However, the widespread use of MCS technology is accompanied by an increasing risk of workers' personal privacy being violated. Typically, workers are required to submit location information in order to participate in task assignments, and the privacy of a worker's location information is a key factor in determining worker participation. Therefore, in order to solve the problem of workers location privacy leakage in the process of task allocation, this paper proposes a location privacy protection method based on local differential privacy (VLDPP). VLDPP constructs a task map based on Voronoi diagram according to the task location, and each task location is mapped into a task area to hide the task location. A local coordinate system is constructed based on the task area, and all workers in the area have their relative location coordinates recalculated and encoded, and then the encoding is perturbed by local differential privacy, thus ensuring workers location privacy. The worker uploads the perturbed location information to the server, which determines the workers in the task by calculating the worker's acceptance rate and completes the task allocation. In addition, this paper improves the availability of perturbed worker location information by dividing the task area at a secondary level. This paper uses the internationally recognized World Check-In Gowalla dataset for experimental evaluation, which shows that the proposed method has good performance in terms of data availability and efficiency, providing adequate privacy guarantees.
computer science, information systems,telecommunications