Differentially Private Coordination of Second-Order Multi-Agent Systems Via Dynamic Encoding-Decoding

This paper considers the privacy protection problems for distributed consensus of discrete-time second-order multiagent systems with limited communication channel capacity and partially measurable state. Differential privacy is applied to protect the sensitive information of agents. We proved that t...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Chinese Control Conference s. 5332 - 5337
Hlavní autori: Zhang, Wenjun, Liang, Yong, Wang, Bing-Chang
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: Technical Committee on Control Theory, Chinese Association of Automation 26.07.2021
Predmet:
Consensus algorithm
Differential privacy
Directed graphs
distributed consensus
Heuristic algorithms
Linear systems
multi-agent systems
Privacy
quantized communication
Visualization
ISSN:1934-1768
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:This paper considers the privacy protection problems for distributed consensus of discrete-time second-order multiagent systems with limited communication channel capacity and partially measurable state. Differential privacy is applied to protect the sensitive information of agents. We proved that two initial state sets of a multi-agent system which satisfies δ-adjacent will export identical information with a certain probability, which means that adversaries will confuse them. In addition, each agent has its real-valued state but can only transmit limited digital signals because of communication constraints. In order to overcome this problem and the condition of partially measurable state, quantized communication method is utilized. Each agent can transmit only integer values to heighbors and estimate the neighbors' real-valued state by dynamic encoding and decoding strategies. A differentially private consensus algorithm based on dynamic encoding-decoding scheme is proposed for discrete-time second-order multi-agent systems. Finally, we built some simulations to visualize our conclusions.
ISSN:1934-1768
DOI:10.23919/CCC52363.2021.9550215