A reliability measure for smart surveillance systems

TL;DR

This paper introduces a new reliability measure for smart surveillance systems that accounts for adversarial threats, using percolation theory and customizable cost functions, demonstrated on a digital twin of a smart building.

Contribution

It generalizes existing reliability measures by incorporating customizable cost functions and applies percolation theory to model smart surveillance system reliability.

Findings

Effective modeling of diverse scenarios with customizable costs

Application to a digital twin of a smart building

Highlights challenges in parameter estimation

Abstract

We present a reliability measure for smart surveillance systems, taking into account the adversarial nature of intrusion. Our approach is based on percolation theory and is a generalisation of Hamedmoghadam et al.'s reliability measure. Specifically, our approach incorporates a customisable cost function to allow modelling a diverse range of situations, such as access restrictions, monitoring, and failures. We demonstrate our approach by applying it to a digital twin of a smart building. However, challenges remain in estimating and modelling the key parameters needed.