British Transport Police go routine BVLOS. Seven boxes, 20,000 miles.

On 24 September, British Transport Police became the first UK force to fly beyond visual line of sight as a routine part of day-to-day policing. It isn’t running a trial or holding a one-off waiver. Its drones launch from drone-in-a-box stations beside the railway. Pilots fly them from BTP’s operations centre at London Bridge, and they carry command-and-control links and live UHD video over BT’s Drone SIM on the EE network.

The numbers explain why. BTP polices more than 20,000 miles of railway, and Dronewatch cites more than 19,300 trespass incidents in 2023/24, which works out to about one every half hour. Most trespass, welfare and disruption calls happen where no officer is nearby. A drone that reaches the scene before anyone else can tell control whether they’re dealing with a person in crisis, kids on the line or a false alarm. That decides whether trains stop, and for how long. Seven docks are live now. BTP plans 14 more over the next two years.

The headline is the aircraft. The hard part is the network.

A trial with one dock is really a flying programme. A network of 21 docks spread across Great Britain is an estate to maintain. Every box is a small unattended building with a charger, batteries, a climate system, a weather station, a door mechanism and a cellular link. Every box sits in a lineside spot chosen for where incidents happen, not for how easy it is to reach. Nobody walks past these boxes.

That’s the core risk for a first-responder model. The promise is that the drone gets there first. If a dock is quietly unavailable because a battery has degraded, the enclosure is running hot or the uplink is marginal, the promise fails without warning. The failure shows up during an incident, not before one.

Three things that matter at 21 docks

Readiness beats flight hours. A police BVLOS network isn’t judged by how much it flies. It’s judged by whether each dock can launch within seconds of a call, at any hour. So the key number is readiness per dock over time, and the patterns that come before a drop in it. Climate faults, thermal alarms and slowly falling battery health all give warning signs days or weeks before a dock goes offline.

Batteries set the budget. On an emergency-response network, flight frequency follows incidents, not a schedule. Batteries at busy docks will age much faster than those at quiet ones. If you track cycles and health per pack, you can replace them on evidence rather than on a calendar or after a failure.

Every flight is evidence. A police drone’s record is not just an ops log. Footage and telemetry may end up in court, in a complaint or in an accident investigation, and a BVLOS approval comes with ongoing obligations to show the operation stays inside its safety case. A clean chain from dock state to flight log to video isn’t optional here.

Why this matters beyond policing

Rail is almost the ideal BVLOS case. The corridors are long and linear, the network is owned by one body, the incident volume is high and every minute of delay has a clear cost. If BTP shows that a national dock network can stay ready over years, not just launch day, the same model can carry over to asset inspection, utilities, highways and ports. Europe is watching. Dronewatch reports that Dutch rail operator ProRail is exploring a similar approach but would need regulatory changes to do it, and Swedish police are building their own record under round-the-clock BVLOS.

Going from 7 to 21 docks isn’t three times the work in a straight line. Every added site brings its own weather, access problems and failure modes. The UK has spent years building the regulatory route to routine BVLOS, and this launch shows the route works. The next two years will show whether the operations can scale with it.

A drone that gets there first is a strong capability. A network that can guarantee it gets there first is what earns public trust in routine BVLOS. So treat the docks as critical infrastructure from day one: monitor readiness continuously, alert on the weak signals before the hard failures, manage batteries on real data, and keep a record that can stand up to scrutiny.