文献类型：期刊文献

英文题名：Localized Privacy Preservation by Innovation Perturbation in a Cooperative LQG Control System

作者：Sheng, Wenliang[1]; Zhao, Zhiyun[1]; Yang, Wen[1]; Yang, Chao[1]

机构：[1] East China University of Science and Technology, Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, Dept. of Automation, Shanghai, 200237, China

年份：2023

起止页码：3486

外文期刊名：Proceedings of the IEEE Conference on Decision and Control

收录：EI(收录号：20240715547433)

语种：英文

摘要：We consider a cooperative Linear Quadratic Gaussian (LQG) control system, in which an individual user owns a local plant whose control inputs are provided by a server. In the cooperation, the user takes the plant states as private information and desires to maximize the privacy preservation while ensuring that the server still provides a certain level of control performance. Moreover, the user requires a privacy scheme that is used locally and is unknown to the server, so that it can create a deviation in the server's knowledge of the states from the true value. To achieve this, we propose two privacy schemes localized at the user side, which inject perturbations in the innovation data sent to the server. For both schemes, firstly, we analyze the privacy preservation quality provided by the scheme and the performance loss in the LQG control caused by it. Secondly, based on the trade-off between them, we propose an optimization problem. Thirdly, we propose a recovery procedure by which the control performance is recovered to the optimal one, i.e., the privacy preservation is achieved without any performance loss in control. Finally, simulations are provided, and we give discussions on the two schemes based on the simulation results. ? 2023 IEEE.