Privacy-Preserving Distributed Precoder Design for Decentralized Estimation

Abstract

We study privacy-preserving precoder design for decentralized estimation in wireless sensor networks where the sensor nodes want their local information such as the channel state information, observation matrices, and observation covariance matrices to be private. We propose a distributed algorithm with closed form expressions to design the precoders and fusion rule that minimize the estimation error by exchanging messages which do not reveal the local information. We derive the privacy limits offered by the proposed algorithm and prove that the algorithm is privacy-preserving. Simulation results illustrate the trade-off between privacy and estimation accuracy of the proposed algorithm.

Distributed Ridge Regression with Feature Partitioning

“…where η > 0 is the regularization parameter. From the normal equation associated with (1), the centralized estimate is given by βo…”

“…Hence, we need algorithms that are capable of processing data spread across multiple agents. They ought to operate in a distributed fashion relying only on the available local information [1]- [10].…”

Joint Transmit and Reflective Beamformer Design for Secure Estimation in IRS-Aided WSNs

“…Traditional cryptographic security is resource-intensive for use in low-power WSNs [1,4,5]. Therefore, physical layer security have been proposed as an low-complexity alternative for secure distributed detection [6][7][8] and secure remote state estimation [9][10][11] in WSNs.…”

Consensus-based Distributed Total Least-squares Estimation Using Parametric Semidefinite Programming

“…Furthermore, collecting all the data in a fusion center creates a single point of failure. Therefore, it is imperative to develop algorithms that are capable of processing data spread across multiple agents [1][2][3][4][5][6][7].…”