Distributed secure estimation for interconnected systems against eavesdropping under energy harvesting constrained encryption.

Abstract

This paper studies the distributed secure estimation problem for interconnected systems, where multiple eavesdroppers intercept transmitted measurement data and collaborate to rebuild states of subsystems. An encryption-decryption model is proposed to achieve resource-efficient privacy protection. In this model, sensors equipped with energy harvesters encrypt measurements before transmission provided that energy is available. The energy-based encryption matrices and estimator parameters are jointly devised to guarantee that estimation error covariances remain bounded for users while deteriorating the estimation performance of eavesdroppers. The influence of the energy level and encryption matrices on the estimation performance of users and eavesdroppers is also reflected. A numerical simulation regarding multi-target tracking is provided to validate the effectiveness of the main results.