Skip to content

Metro and urban rail

Knowing where your people are when something happens on a platform

This is the case the system was designed for, and the one that best exploits what makes it different: a narrow, heavily constrained and perfectly known topology.

The problem as it is actually lived

An incident in a station is resolved by radio and by whoever remembers the shift roster. The control room knows which teams are on duty, but not where each of them is right now: whether the nearest team is in the concourse or two levels down, whether they are coming up the east or west stairs, whether the responder really is the one who will get there fastest.

The consequence is rarely dramatic, but it is constant: minutes of coordination lost asking where people are. And when the incident is serious, those minutes are exactly the margin you do not have.

Then comes the second half of the problem: reconstructing what happened. Who arrived first, how long it took, which way they came in. Without data, it is one person’s word against another’s.

Half the shift happens inside a train

On a metro network, patrol work does not happen only on the platform: a large share of it is on board, between stations. And that is exactly where the officer is most alone and where an incident is hardest to reach.

The on-board layer fits one BLE beacon per carriage and turns that blind stretch into exact information: which train, which car and which station it is at. It stops being a coordinate with an error margin and becomes an identification.

The operational change is concrete: if you know the officer is in car 3 of the service arriving in two minutes, you can go and meet them on the right platform, at the right door instead of searching.

The beacons sit inside the rolling stock, are fitted during a depot inspection and require touching no platform and no tunnel. It is optional: the base layer works on its own. Detail in on-board tracking.

Where it fits in operations

ProcessHow it integratesOutcome
Security and patrollingApp on duty terminals, visualisation in the portalReliable personnel position and coordination during incidents
Control roomAPI feeding positions and events into the incident manager and operational mapIncident records enriched with location and trajectory
Service planningHistory export and coverage maps by zone, shift and — with the on-board layer — by train and carShifts and patrol routes tuned to real coverage, and evidence of the coverage contracted
Occupational risk preventionAudit trail of events and alerts, anonymised where appropriateObjective evidence for incident investigation
Systems and cybersecurityAPI integration with TLS, 2FA, roles and ENS-compliant loggingCompliance and risk control without opening surface
Procurement and assetsReuse of the existing Wi-Fi network, no civil worksContained cost and fast adoption

Why metro is the best case

There is a specific technical reason, and it is worth understanding because it explains the performance gap against other indoor settings.

  • The topology is narrow and heavily constrained. Platform, concourse, corridor, stairs, tunnel. Almost all the volume of a station is space you cannot walk through, and that is usable information: map matching rules out enormous numbers of candidate positions.
  • The topology is stable. A station does not change shape. What you calibrate today still holds tomorrow.
  • The Wi-Fi is already installed. Metro networks have been rolling out station coverage for years for other reasons. The project’s infrastructure is already paid for.
  • The use case is safety, not fine accuracy. Nobody needs centimetres. You need to know which platform and which entrance.

What gets measured

  • Mean time to locate the nearest team during an incident
  • Share of trains with evidenced presence per shift, where the on-board layer is deployed
  • Position publication latency
  • Mean accuracy per station, checked against reference points
  • Share of incidents with full traceability of the response

These metrics are agreed during the pilot before it starts, with numeric thresholds.

Shall we look at one specific line?

Pick the line with the most incidents, or the one that worries you most. That is where it makes sense to start.

We reply within one business day.