文献类型：期刊文献

英文题名：Privacy preservation based on reinforcement learning in cooperative LQG control systems

作者：Lin, Wenhao[1];Yang, Wen[1];Yang, Chao[1]

机构：[1]East China Univ Sci & Technol, Dept Automat, Key Lab Smart Mfg Energy Chem Proc, Minist Educ, Shanghai 200237, Peoples R China

年份：2026

卷号：216

外文期刊名：SYSTEMS & CONTROL LETTERS

收录：;EI(收录号：20262721023383);WOS:【SCI-EXPANDED(收录号:WOS:001815004300001)】；

基金：This work is supported in part by the National Key R&D Program of China (2023YFF1204805) , National Natural Science Foundation of China under Grant (62336005) .

语种：英文

外文关键词：Privacy preservation; Reinforcement learning; Cooperative networked control systems

摘要：This letter focuses on the problem of privacy preservation in a cooperative linear quadratic Gaussian (LQG) control system with a single user and a server. In this letter, we study the privacy preservation problem for the user during the data transmission process, and we propose a scheme of adding intermittent noise at certain moments while being constrained by the control cost and resource consumption. The problem under consideration is a multi-objective optimization designed to achieve comprehensive optimal performance. To address this problem, we formulate an associated Markov decision process (MDP) and propose an algorithm based on reinforcement learning (RL) technique to solve it. Finally, we present numerical examples to demonstrate the effectiveness of the proposed algorithm.