Preprint: Anj Simmons, A reliability measure for smart surveillance systems. https://arxiv.org/abs/2202.09339
The ontology in this repository extends Microsoft's opendigitaltwins-building ontology with additional classes and properties for modelling and reasoning about surveillance and security. In particular, it adds the availability
property to support reasoning about how likely a surveillance system is to actually be functioning at the time of an intrusion, as well as the concept of failclosed
(e.g. a safe door that locks shut if there are any faults) and failopen
(e.g. emergency exits that swing open if the building loses power).
The sample directory provides a sample digital twin constructed using the ontology. Both a spreadsheet representation and .json
export are provided. This can be imported into the Azure Digital Twins Explorer tool (after loading the opendigitaltwins-building models followed by the A2I2 models in the ontology directory).
The scripts directory contains the code to compute a reliability measure, α, of the surveillance system described by the digital twin, taking into account the adversarial nature of intrusion.
The analysis script is implemented in Python3. There is also an optional visualisation of budget vs α implemented in R. To install all dependencies in a conda environment run:
conda create -n surveillance python=3 r-essentials r-base
conda activate surveillance
cd Scripts
./analyse.sh