.png){kind=icon}
.png){kind=icon}
Return
Privacy-preserving distributed estimation for interconnected dynamic systems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.automatica.2025.112277.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper investigates the problem of privacy protection in distributed estimation for interconnected dynamic systems. The exchange of information between subsystems during weighted sum aggregation poses significant privacy risks to distributed estimation. To address these concerns, we propose a noise contamination mechanism for private matrix-vector multiplication and private weighted sum aggregation. This mechanism decomposes the matrix into two column-rank-deficient matrices and injects noises derived from their null spaces to protect the confidentiality of the vector during the matrix-vector multiplication. Several conditions are established to ensure that the mechanism achieves complete confidentiality. Subsequently, we design a double-layer encryption scheme for private weighted sum aggregation with hidden weights, incorporating private matrix-vector multiplication and the Paillier cryptosystem. Furthermore, we develop privacy-preserving distributed estimators for interconnected dynamic systems using the proposed aggregation schemes. Our analysis demonstrates that both subsystems and their distributed estimators achieve effective privacy protection. Finally, an illustrative example is provided to validate the effectiveness of the proposed methods. (c) 2025 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
Keywords:
Privacy-preserving distributed estimator
Interconnected dynamic system
Private matrix-vector multiplication
Private weighted sum aggregation
Paillier cryptosystem