近期，实验室博士生胡仲瑞作为第一作者，导师吴立刚教授作为通讯作者的论文“Preserving State and Control Privacies in Networked Systems with Tokenized Polytopic Transforms”已被国际权威期刊IEEE Transactions on Circuits and Systems II: Express Briefs录用。

该论文针对网络控制系统中一类具有具有状态与输入约束的线性被控对象，提出了基于整数标识符映射结合多面体仿射变换的隐私保护控制策略。该策略的设计利用了鲁棒模型预测控制律在状态约束集合上分段仿射的性质，得到了一个包含子区域与虚拟子区域的多面体集合以及一个包含上述多面体的放射变换集合，并给出了这两个集合间基于整数标识符的映射关系。已有相关成果中主要考虑系统状态的隐私保护。在本文中，当考虑攻击者(或窃听者)存在于控制器端，并能够获取控制器的输入、输出以及与控制律相关信息的情形下，该策略保证了系统的状态和控制输入的语义安全。

Abstract

This brief studies the problem of preserving state and control privacies in a networked system where the plant is a constrained discrete-time linear system with bounded disturbances. Our approach to this problem is a privacy-preserving control scheme which assigns integer tokens to the combination of affine transforms and qualifying polytopes within state constraints. Specifically, we utilize the piecewise affine property of a feedback control law obtained with the explicit model predictive control technique and design tokenized affine transforms for the control law and the polytopic partition of its domain. The qualifying polytopes are referred to as subregions and virtual subregions which are generated within the partitioned domain mentioned above. We show that the proposed scheme is semantically secure against privacy extraction by an eavesdropper. Simulation results demonstrate the validity of the proposed scheme in preserving state and control privacies.