University of MaineElectrical and Computer Engineering

CAST-Map a living map of every connected device on the factory floor, and what to worry about first

IEEE International Conference on Consumer ElectronicsICCE 2025
CAST-Map scorecard
A live inventory, re-scanned on a schedule. Each device is scored by what it must protect, then weighted by what has happened to its look-alikes.
Latest scan
–
0 devices 
Look-alike groups and their weight
Facility criticality
Per-device scoring
CAST-Map
group-weighted
What the paper built and tested
Live
Tracked a real additive manufacturing testbed at UMaine: hardware, software, ports, USB devices, links, re-scanned on a schedule and encrypted in transit.
No agent
Printers and controllers are never touched. CAST-Map maps them from the network, so sensitive machines stay untouched and still show up on the map.
2,000
sub-systems in the largest simulated facility, scored automatically. Affinity Propagation gave the most stable facility scores across sizes.
Scores in the paper come from randomized facilities modeled on the tracked testbed. The cell shown here is an illustration of the method, not a measured result, and the digital twin step shows what the map enables rather than an experiment reported in the paper. The lesson carries: keep a living map, let similar devices share the lessons of an incident, and rehearse attacks on the twin.
Funded by the U.S. Army Engineer Research and Development Center
For more details, contact
Nickolas Millett, lead student researcher nickolas.millett@maine.edu·Prabuddha Chakraborty, SIEGE Lab PI prabuddha@maine.edu