Cooperative Differentially Private LQG Control With Measurement Aggregation

Article Automation & Control Systems

Differentially Private LQ Control

Kasra Yazdani et al.

Summary: In order to protect sensitive data while allowing system operation, new approaches are needed for the increasing number of multi-agent systems. This article presents a private multiagent LQ control framework that uses differential privacy to protect agents' state trajectories. The impact of privacy is quantified in terms of the information shared, the cost of privacy in control theory, and the tradeoffs between privacy and performance. Analysis is conducted using conventional control-theoretic terms and guidelines for calibrating privacy as a function of system parameters are developed. Numerical results show that system performance remains desirable even under strict privacy requirements.

IEEE TRANSACTIONS ON AUTOMATIC CONTROL (2023)

Add to Collection

Article Automation & Control Systems

Private Weighted Sum Aggregation

Andreea B. Alexandru et al.

Summary: This article addresses the problem of private weighted sum aggregation and proposes different secure multiparty computation schemes to achieve the private aggregation of weights.

IEEE TRANSACTIONS ON CONTROL OF NETWORK SYSTEMS (2022)

Add to Collection

Article Automation & Control Systems

Modular control under privacy protection: Fundamental trade-offs

Yu Kawano et al.

Summary: In privacy-preserving controller design, there is a trade-off between privacy level and control performances. Each module can increase the privacy level of its reference signal by tuning local controller parameters, but there exists a lower bound on the differential privacy level when noise distribution is fixed.

AUTOMATICA (2021)

Add to Collection

Article Automation & Control Systems