Optimizing encrypted control algorithms for real-time secure control

Abstract

In network control systems (NCSs), one obvious advantage of using homomorphic encryption is to protect data privacy while remotely computing the control function without decryption. Homomorphism analysis shows that it is more secure to protect the control system with fully homomorphic encryption (FHE) than with partially homomorphic encryption (PHE). However, the high computational complexity of FHE hinders its acceptance in NCSs. In this paper, FHE is utilized to highly enhance the security of NCSs against potential eavesdropping attacks. To guarantee control efficiency, an optimization scheme of the encrypted control algorithm is designed to make FHE practically applicable for real-time control. Besides, the secure controller achieves the expected control performance and inherits asymptotic stability. The experimental results can demonstrate the effectiveness of the proposed encrypted control system in terms of efficiency, security, and stability.