Drone-in-a-box: what one dock actually covers in a day

A drone-in-a-box installation is a weatherproof enclosure holding a drone, a charging system and enough connectivity to take a command and return a video stream. On a trigger — a schedule, a sensor, a person in a control room — the roof opens, the aircraft launches, flies a stored route and lands back on the pad.

The part that gets left out of that description is that the resource is singular. One dock holds one aircraft. While that aircraft is airborne, the dock is empty and cannot receive anything else. While it is charging, the dock is occupied and cannot launch. There is no queue and no overlap. Sorties from a single dock are strictly serial.

That sounds obvious written down, and it is routinely missed in planning, because the demonstration everyone sees is a single flight. A single flight always looks like abundant capacity. The question that decides the business case is what the tenth flight of the day looks like, and whether it can happen at the moment you need it.

Take the most widely deployed reference point. For the Dock 2, DJI states that real-time blade detection, safety checks and take-off complete in 45 seconds, given a robust network signal. DJI also states a charging time of 32 minutes to bring the aircraft from 20 to 90 per cent in a 25 °C environment.

Both figures are useful and both carry a qualifier that belongs in the planning rather than in the brochure. The 45 seconds assumes the network is good; a dock at the far corner of a site on marginal coverage is not a 45-second dock, and the checks that make up those 45 seconds are the ones you would not want skipped. The 32 minutes assumes 25 °C and a partial charge band — from 20 per cent, not from empty, and to 90 per cent, not to full.

Neither qualifier is a criticism of the product. They are the ordinary conditions under which a specification is measured, and quoting the figures without them is what produces disappointed operators. The right use of these two numbers is as inputs to an arithmetic, which is what the next section does.

The planning unit is the cycle: launch, fly, land, charge, ready. Launch is negligible at three quarters of a minute. Charge, on the published figure, is a little over half an hour under favourable conditions. So the cycle is roughly the flight time plus half an hour.

Put your own flight time in and the arithmetic does itself. A ten-minute perimeter lap on a cycle of about forty minutes gives something on the order of one lap every forty minutes, and rather fewer than that across a real day once weather holds, maintenance windows and scheduled inspection routes are taken out. A longer inspection sortie stretches the flight side and shortens the number of cycles further.

This is why the honest comparison is never "drone versus guard" but "how many covered laps do I get, at what interval". If the requirement is a lap every twenty minutes, one dock does not meet it at any flight duration, because the charge alone exceeds the interval. The answer then is a second dock, a different cadence, or a different technology — and that is a decision far better made before procurement than after.

Scheduled rounds are the easy case, because you control the timing. The hard case is event-driven use, where the dock exists to put eyes on an alarm quickly.

Here the serial constraint bites in a way that scheduling cannot smooth. An alarm comes in, the aircraft launches, verifies, returns and starts charging. For the next half hour or so, that dock cannot answer anything. If a second alarm arrives in that window — and on a site where alarms cluster, which is most sites, that is not a remote possibility — the response is whatever it was before you installed the dock.

So the specification question is not "how fast does it respond" but "what is the availability across the alarm pattern we actually have". That requires looking at your own alarm log for a year: how often two alarms fall within thirty minutes, and what the consequence of the second one going unverified would be. Operators who do that arrive at a number of docks. Operators who do not arrive at one dock and a disappointment. The wider set of things that break once a perimeter programme grows past its first site is covered in what breaks after the pilot.

It is easy to read the box as weather protection: something that keeps rain off an expensive aircraft. It does that, and it does something less obvious that matters more to the cycle.

Recall that the 32-minute charge figure is quoted at 25 °C. Charging chemistry is temperature-dependent, and the published figure is a warm-conditions figure. A dock standing on a yard in northern Europe in February is not working in a 25 °C environment unless the enclosure is actively bringing it there — which is one of the things a conditioned enclosure is for, and one of the reasons a dock draws mains power rather than running off the aircraft battery.

The practical consequence is that the cycle you planned in a summer trial is not the cycle you get in winter, and the gap is not a fault. It is the difference between a specification condition and a site condition. Ask for the charge figure at the temperature range your site actually sees, and ask what the enclosure does to close that gap. If the answer is that the enclosure is passive, plan a longer winter cycle and say so in the shift plan rather than discovering it in January.

A dock is a fixed installation. It needs a footing, mains power, network, and a piece of ground whose owner has agreed to it. Once it is there, its position sets the geometry of everything it can do: what is within range, what is within the return-to-home envelope, what the aircraft would overfly on any given route, and what lies beneath the volume it needs in a failure case.

Those are the same questions any authorisation asks, and they get asked once for a fixed installation rather than repeatedly. That is the genuine advantage of the box and the reason it suits a site that is covered over and over: the assessment is done once and stays done.

It is also the reason siting is not a facilities decision to be settled after the aircraft is chosen. Moving a dock fifty metres can change which routes are assessable. Choosing the aircraft first and the position afterwards inverts the dependency, and the cost of that inversion appears late, when the position is already poured.

The box answers one shape of problem very well: the same ground, covered repeatedly, on a cadence no one wants to staff. Where a single point is visited often enough, a fixed installation amortises its siting, its power, its network and its one-time assessment across hundreds of sorties.

Where that is not true — where the requirement is a different location every day and no point is visited often enough — the installation never amortises, and the answer is a crew that arrives with the aircraft in a vehicle. That is a different operating model with different regulatory questions, and we have set it out separately in launching drones from a vehicle. The third case, continuous presence over one fixed point rather than repeated sorties, is the tether, covered in tethered drones.

Three models, three shapes of requirement. The mistake worth avoiding is buying the one with the best demonstration rather than the one matching your cadence. Describe the cadence first — how often, over what ground, at what interval, with what response time on an event — and the model chooses itself. The German companion to this article, which covers the same technology from the site-security side, is Drone-in-a-Box: Automatisierte Drohnen für Werksgelände.