Privacy-Preserving Adaptive Consensus based Cubature Kalman Filter for Distributed Sensor Networks

Abstract

Information security is an unsolved problem of existing distributed state estimation. In this paper, a privacy-preserving adaptive consensus-based cubature Kalman filter (PAC-CKF) with certain estimation accuracy and convergence speed is proposed to improve the information security of distributed sensor networks. By combining the state-decomposition mechanism with adaptive average consensus in the frame of cubature Kalman filter, the proposed algorithm can ensure both the network security and estimation accuracy under limited consensus iterations. Simulations are performed to demonstrate the effectiveness of estimation accuracy, privacy preservation, and convergence rate of the proposed algorithm.