A differentially private square root unscented Kalman filter for protecting process parameters in ICPSs

Area of Science:

• Cyber-Physical Systems
• Data Privacy
• Control Systems

Background:

• Industrial cyber-physical systems (ICPSs) face challenges in balancing data privacy and utility.
• Existing methods struggle to effectively protect process parameters without sacrificing data usefulness.

Purpose of the Study:

• To develop advanced differential privacy algorithms for ICPSs.
• To enhance the privacy of process parameters while maintaining their utility.
• To address the privacy-utility trade-off in ICPS data.

Main Methods:

• Derivation of Kalman filter-based differential privacy algorithms.
• Development of an unscented Kalman filter-based differential privacy algorithm.
• Proposal of a differentially private square root unscented Kalman filter algorithm.

Main Results:

• The proposed algorithms effectively protect process parameter privacy.
• The differentially private square root unscented Kalman filter algorithm improves data utility.
• Experimental validation on a numerical control lathe demonstrates effectiveness.

Conclusions:

• The novel differential privacy algorithms successfully address the privacy-utility trade-off in ICPSs.
• The differentially private square root unscented Kalman filter offers a promising solution for secure and useful process data.
• The findings contribute to more secure and efficient industrial cyber-physical systems.