一种基于矩阵补全的智能电网数据隐私保护方案  

A matrix completion based data privacy protection scheme in smart grid

作　　者：王德庆 许建[1] WANG Deqing;XU Jian(School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China)

机构地区：[1]南京邮电大学计算机学院,江苏南京210023

出　　处：《南京邮电大学学报（自然科学版）》2022年第2期103-110,共8页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

基　　金：中国博士后科学基金(2019M651922)资助项目。

摘　　要：针对智能电网中不同信任域间的数据隐私保护问题,提出了一种基于矩阵补全的智能电网数据隐私保护方案。该方案从少量节点产生的模糊值子集中估计出原始数据相关子空间,通过矩阵补全来修复相关矩阵中的缺失元素。为保证数据隐私,通过添加一种与原始数据具有相同统计特性的噪声来增加随机扰动。仿真结果表明,该方案具有较好的隐私保护效果和数据可用性,同时能够有效降低智能电网中的通信开销。To protect data privacy among different trust domains of the smart grid,a matrix completion based data privacy protection scheme is proposed for smart grids.The scheme estimates the original data correlation subspace through a subset of obfuscated values generated by a small number of nodes,and repairs the missing elements in the correlation matrix via matrix completion.Further,to protect data privacy,the specific noises that have the same statistical properties as the original data have been added to increase random perturbation.Simulation results demonstrate that the proposed scheme achieves a good performance from the perspective of data availability and privacy protection,and it can also reduce network communication overhead effectively.

关 键 词：智能电网 隐私保护 矩阵补全 数据扰动 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]