Privacy Preserving Controller Synthesis via Belief Abstraction

TL;DR

This paper introduces a belief-based privacy notion for systems, proposing belief space abstraction and controller synthesis methods to ensure privacy, demonstrated through an illustrative example.

Contribution

It presents a novel belief abstraction technique with efficient algorithms and two controller synthesis approaches for privacy preservation in systems.

Findings

Belief space dynamics are shown to be mixed monotone.

Efficient abstraction algorithms are developed.

Two controller synthesis methods successfully preserve privacy.

Abstract

Privacy is a crucial concern in many systems in addition to their given tasks. We consider a new notion of privacy based on beliefs of the system states, which is closely related to opacity in discrete event systems. To guarantee the privacy requirement, we propose to abstract the belief space whose dynamics is shown to be mixed monotone where efficient abstraction algorithm exists. Based on the abstraction, we propose two different approaches to synthesize controllers of the system to preserve privacy with an illustrative example.