文献类型：期刊文献

中文题名：协作性远程状态估计中的隐私保护研究

英文题名：Privacy Preservation in Cooperative Remote State Estimation

作者：于小平[1,2];杨超[1,2]

机构：[1]华东理工大学能源化工过程智能制造教育部重点实验室,上海200237;[2]华东理工大学信息科学与工程学院,上海200237

年份：2022

卷号：29

期号：6

起止页码：996

中文期刊名：控制工程

外文期刊名：Control Engineering of China

收录：CSTPCD;;北大核心:【北大核心2020】；CSCD:【CSCD_E2021_2022】；

基金：上海市自然科学基金面上项目(19ZR1413500)。

语种：中文

中文关键词：卡尔曼滤波器;隐私机制;隐私保护;凸优化

外文关键词：Kalman filter;privacy mechanism;privacy preservation;convex optimization

摘要：网络化控制系统中的隐私保护问题是实际应用中的重要问题。考虑一个包含用户方和服务方的远程估计系统,其中用户方向服务方发送系统状态观测值,由服务方为其计算估计值。用户方在合作中存在隐私保护的需求,希望服务方产生具有偏移量的估计值。针对此需求,首先,提出用户方本地注入噪声的隐私保护机制;其次,选取原始信息和采取保护机制后信息之间的偏移量的方差作为隐私衡量指标;在此基础上,进一步分析得到了隐私衡量指标和估计质量的迭代计算公式以及两者之间的关系;最后,建立了隐私保护水平和估计效用之间的权衡问题,并利用凸优化方法解决了所提出的问题。
Privacy preservation in networked control systems is an important problem in practical application.We consider a remote estimation system consisting of a user and a server,in which the user sends its measurements of the system states to the server and the server calculates the estimates for the user.The user demands privacy preservation in the cooperation and expects the server to produce deviated estimates from the true ones.To meet this requirement,firstly,a privacy preservation mechanism of adding noises locally adopted by the user is proposed.Secondly,a novel privacy metric is proposed,which is the covariance of the deviation between the original information and the information under the preservation mechanism.Based on these,the iterative calculation of privacy metric and estimation performance and the relationship between them are obtained by further analysis.Finally,the trade-off problem between privacy preservation level and estimation performance is proposed,which can be solved by convex optimization.