I recently finished Daniel Suarez’s excellent thriller Kill Decision. The major plot point: parties unknown have been releasing autonomous, armed drones that are killing people in a variety of ways. The drones are capable of insect-level intelligence and swarming behavior, and of autonomously finding human targets and bombing or shooting them. Suarez asks a fairly provocative question: would America’s love affair with drones change if other countries, or criminal syndicates, or even individuals had them and used them as freely in the US as we use them elsewhere? Great plot, well-written, and solid characterizations– by far the best of his books so far. Highly recommended.
Anyway, with that in mind, I saw an article on the Lawfare blog about a guy who equipped an inexpensive commercial drone with a paintball marker. This video shows it in action, hitting targets easily while maneuvering slowly. The video’s a little fear-monger-y, but the narrator is right: “it seems inevitable” that these drones will be used in ways the manufacturer didn’t anticipate. I sent the video to a couple of coworkers, one of whom asked “I wonder how hard it is to shoot accurately with it?” That got me to thinking… so off the top of my head, I jotted down a few factors that would affect the accuracy of a firearm-equipped drone. Note that here I’m talking about an autonomous UAV, not a remotely-piloted, man-in-the-loop drone.
- What’s it for? What kind of range and endurance do you need? It would be easy to build a sort of launch rack that would launch a drone to check out a target that triggered a tripwire, motion detector, etc. It’d be a little harder to build one that could autonomously navigate, but definitely doable– as Paul proved with his Charlie-following project. See also: the Burrito Bomber, which can follow waypoints and then deliver a payload on target. Drones to sneak into somewhere and snipe a single target would have different range/payload requirements than a patrol or incident-reponse drone. This drives the weight of the drone (since more range requires more fuel).
- What’s it packing? The purpose of the drone dictates what kind of firearm you want it to carry. Some of Suarez’s drones had short-barrelled .38 pistols, which are plenty good enough to kill from close range but wouldn’t be very accurate past around 35 feet or so. A longer barrel and a heavier round would provide better accuracy, at the cost of weight and size.
- How much range do you need? A sniper drone that can shoot targets from 1500yds is definitely feasible— use a .50 Barrett, for example. It would be heavy and range-limited, though, unless you wanted to make it bigger. In general, heavier bullets are more stable and give you better accuracy, but they’re heavier to carry and shoot.
- How stable is the drone? A light drone that’s sensitive to wind, etc. will be harder-pressed to make accurate shots. Gyrostabilizing the gun platform would help, but it would add a weight and cost penalty (including for power for the gyros, plus the gyros themselves). The bigger the drone, the more sensors, power, and ammo you can carry… but the more noise, infrared, and visual signature it creates. A small sneaky drone may be a better deal than a large, more powerful one.
- What can you see? In other words, what kind of sensors do you have for aiming? How good is their resolution and range? Do they have to be automated? If so, you need to be able to either fire at the centroid of the target or track interesting parts, like wheels of a truck or a person’s head), using machine vision.
- Where are you pointing the gun, and how accurate can you be? What kind of angular resolution does the gun-pointing system have? If you’re willing to slow to a dead hover, or nearly so, you can be very accurate (as in the video above). If you want to go faster, you’ll have a more challenging set of requirements– you have to be able to point the gun while the drone’s moving, and changing its aim point means fighting inertia in a way you don’t have to worry about in a hover.
There are lots of other more subtle considerations, I’m sure; these are just what I came up with in 5 minutes. Any engineer, pilot, or armorer could come up with a couple dozen more without too much effort. Of course, you could just buy a pre made system like this one from Autocopter. Isn’t it great to know they’ll lease you as many UAVs as you need? Just for a ballpark figure, Autocopter quotes an 8Kg payload on their smallest drones– figure 3Kg for a cut-down M4 and that leaves you a reasonable 5Kg for sensors, guidance, navigation, and control.
What could you do with such drones? The mind boggles. Imagine that, say, your favorite Mexican drug cartel cooked up a bunch of these in their machine shops and used them to guard the pot farms they run in national forests. Or say the white-supremacy militia guys in Idaho built some for sovereign defense. Or suppose you built 100 or so of them, staged them inside an empty 18-wheeler with a tarp over the top, then launched them into Candlestick Park during a 49ers game. There are all sorts of movie-plot-worthy applications for these drones, to say nothing of the ones Suarez wrote about.
Meanwhile, the February 2013 NASA Aviation Safety Reporting System (ASRS) newsletter is full of safety reports filed after drones got into airspace where they weren’t supposed to be… and these were piloted, unarmed drones. How careful do you think these hypothetical armed drones would be about respecting the National Airspace System? I think I’ll be extra careful when flying around… that smudge on the windscreen might turn out to be an armed autonomous drone.