Encrypted Observer-based Control for Linear Continuous-Time Systems

Hung Nguyen,Binh Nguyen,Hyung-Gohn Lee,Hyo-Sung Ahn
DOI: https://doi.org/10.48550/arXiv.2303.00963
2023-03-02
Abstract:This paper is concerned with the stability analysis of encrypted observer-based control for linear continuous-time systems. Since conventional encryption has limited ability to deploy in continuous-time integral computation, our work presents systematically a new design of encryption for a continuous-time observer-based control scheme. To be specific, in this paper, both control parameters and signals are encrypted by the learning-with-errors (LWE) encryption to avoid data eavesdropping. Furthermore, we propose encrypted computations for the observer-based controller based on its discrete-time model, and present a continuous-time virtual dynamics of the controller for further stability analysis. Accordingly, we present novel stability criteria by introducing linear matrix inequalities (LMIs)-based conditions associated with quantization gains and sampling intervals. The established stability criteria with theoretical proofs based on a discontinuous Lyapunov functional possibly provide a way to select quantization gains and sampling intervals to guarantee the stability of the closed-loop system. Numerical results on DC motor control corresponding to several quantization gains and sampling intervals demonstrate the validity of our method.
Systems and Control
What problem does this paper attempt to address?