Privacy Preserving for Switched Systems Under Robust Data-Driven Predictive Control
针对切换系统中的强非线性、不确定性和数据隐私挑战，提出一种隐私保护鲁棒无模型自适应预测控制方法，通过设计性能依赖的拉普拉斯噪声平衡系统性能与隐私，并给出隐私水平分析方法。
Differential privacy preserving ensures the privacy of the system data by adding certain regular noises to the data to cover up the real information. The main challenges of strong nonlinearities, uncertainty, and data privacy are considered together for switched systems, and a novel privacy-preserving robust model-free adaptive predictive control (PPR-MFAPC) method is proposed that guarantees both <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{\infty }$ </tex-math></inline-formula> performance and system privacy. At first, a performance-dependent differential privacy noise conforming the Laplace distribution is designed, which can adaptively adjust the noise size to balance the system performance and privacy. Then, a novel privacy level analysis with evaluation method is presented. Subsequently, the strong uncertainties of switched systems is solved through a dynamic linearization method. On this basis, a novel cost function is designed by considering both the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{\infty }$ </tex-math></inline-formula> performance and system privacy to balance the system performance and privacy from the perspective of control design. Further, by incorporating a parameter estimator and a prediction algorithm, the private MFAPC anti-noise controller is obtained. Finally, the feasibility of the PPR-MFAPC is explained with illustrative example.