2020 · 12 citations · 11 references
ComputingComputer ScienceCybersecurityCyber-physical SystemsMachine LearningOptimizationData PrivacyOptimal ControlStochastic ControlAdditive NoiseFisher InformationLinear Quadratic RegulationLinear Quadratic RegulatorLinear-quadratic ControlPrivacyPrivate OptimizationWorst-case Scenario
This article introduces and solves a new privacy-related optimization problem for cyber-physical systems where an adversary tries to learn the system dynamics. In the context of linear quadratic systems, we consider the problem of achieving a small cost while balancing the need for keeping knowledge about the model's parameters private. To this end, we formulate a Fisher information regularized version of the linear quadratic regulator with cheap cost. Here the control operator is allowed to not only control the plant but also mask its state by injecting further noise. Within the class of linear policies with additive noise, we solve this problem and show that the optimal noise distribution is Gaussian with state dependent covariance. Next, we prove that the optimal linear feedback law is the same as without regularization. Finally, to motivate our proposed scheme, we formulate for scalar systems an equivalent maximin problem for the worst-case scenario in which the adversary has full knowledge of all other inputs and outputs. Here, our policies are maximin optimal with respect to maximizing the variance over all asymptotically unbiased estimators.
11
System identification—Theory for the user
Naresh K. Sinha · Automatica · 1989 · 9.2K citations
Optimization by Vector Space Methods
Students Quarterly Journal · 1970 · 6.1K citations
Natural Sciences, Industrial Engineering, Design Optimization +6
Differentially Private Filtering
Jérôme Le Ny, George J. Pappas · IEEE Transactions on Automatic Control · 2013 · 331 citations
Differential privacy in control and network systems
Jorge Cortés, Geir E. Dullerud, Shuo Han et al. · 2016 · 223 citations