计算机科学
差别隐私
计算机安全
数据收集
范式转换
数据科学
数据挖掘
哲学
统计
数学
认识论
作者
Ting Yuan,Taochun Wang,Yong Qiang,Lei Shen,Fulong Chen,Chuanxin Zhao
标识
DOI:10.1016/j.comnet.2024.110421
摘要
Location information is often required for task allocation in mobile crowdsensing. However, directly uploading location information can raise concerns about privacy leakage among users. To address this issue and protect users privacy when uploading location data to the server, this paper proposes a novel method based on differential privacy for preserving location information. The main idea is to utilize Hilbert mapping to transform the two-dimensional location coordinate of the user into a one-dimensional Hilbert index. Then, Laplace noise is added to the index to introduce a perturbed index, ensuring the protection of location privacy. To enhance the usability of the perturbed index, this paper treats the Laplace noise as a random variable and employs mathematical integration to measure the distance between users and tasks. Leveraging the characteristic of the Hilbert index, wherein a small change in the indexes can result in a large distance between corresponding coordinates, the combination of differential privacy and Hilbert mapping offers enhanced effectiveness compared to conventional differential privacy location preservation schemes. Finally, the proposed scheme is evaluated using the real dataset, and the experiments validate its efficacy.
科研通智能强力驱动
Strongly Powered by AbleSci AI