| 引用本文： |
-
唐聪,薛乔,王箭,等.车联网中基于度量本地差分隐私的ぜ合数据隐私保护机制[J].电讯技术，2026，66（1）： - . [点击复制]
- TANG Cong,XUE Qiao,WANG Jian,et al.A Metric Local Differential Privacy-based Mechanism for Set-valued Data Privacy Protection in the Internet of Vehicles[J].，2026，66（1）： - . [点击复制]
|
|
| 摘要: |
| 现有的针对集合型数据的隐私保护机制如果直接应用到车联网中会对频率分布估计结果的准确度造成很大影响。针对这一不足，提出了一种基于度量本地差分隐私模型的对称差私有集合（Symmetric Difference Private Set，SDPrivSet）协议。该协议中，用户在本地将数据扰动后提交给服务器，服务器则根据接收到的扰动数据估计出真实数据的频率分布。该协议提供了严格的数据隐私保护，在用户端和服务器端具有低计算开销，并且在进行统计分析时具有高数据效用。在真实数据集上的实验结果表明，SDPrivSet协议在任意原始数据域和集合大小以及隐私预算下性能都是最优的，相较于现有协议提升了至少34.20%，并且在集合大小和隐私预算较大时性能提升更为明显。 |
| 关键词: 车联网 度量本地差分隐私 频率估计 集合型数据 隐私保护 |
| DOI：10.20079/j.issn.1001-893x.241127002 |
|
| 基金项目:国家自然科学基金资助项目（62302214）；国家自然科学基金联合基金重点项目（U2433205）；江苏省重点研发计划（产业前瞻与关键核心技术）项目（BE2022068，BE2022068-1）；稳定支持国防特色学科基础研究项目（ILF240061A24）；中国高校产学研创新基金——新一代信息技术创新项目课题（2023IT049） |
|
| A Metric Local Differential Privacy-based Mechanism for Set-valued Data Privacy Protection in the Internet of Vehicles |
| TANG Cong,XUE Qiao,WANG Jian,ZHANG Yan |
| (College of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China) |
| Abstract: |
| Existing privacy protection mechanisms for set-valued data,if directly applied to the Internet of Vehicles(IoV),can significantly impact the accuracy of frequency distribution estimation results.To address this shortcoming,a symmetric difference private set(SDPrivSet) protocol based on the metric local differential privacy model is proposed.In SDPrivSet,users locally perturb data before submitting them to the server,which then estimates the true frequency distribution based on the received perturbed data.This protocol provides strict data privacy protection,has low computational overhead on both the user and server sides,and maintains high data utility during statistical analysis.Experimental results on real datasets show that the SDPrivSet protocol performs optimally under any original data domain,set size,and privacy budget.Compared with existing protocols,it enhances performance by at least 34.20%,with more significant performance improvements when the set size and privacy budget are larger. |
| Key words: Internet of Vehicles metric local differential privacy frequency estimation set-valued data privacy protection |
《电讯技术》编辑部 