Original URL: http://www.theregister.co.uk/2012/09/03/smartphone_app_can_call_in_drone_strikes_not_really/

War-Droid: Smartphone app 'that could CALL IN drone strikes'

Can also be used as police speed gun, golf rangefinder

By Lewis Page

Posted in Government, 3rd September 2012 07:00 GMT

Engineering boffins in Missouri say they have come up with a new Android app which allows any modern smartphone - potentially - to call in a devastating precision airstrike by simply snapping a picture of the target to be struck.

"On the battlefield, a soldier needs a rangefinder, compass, GPS and other tools to do reconnaissance before calling in an air strike," explains Qia Wang, the doctoral student leading work on the war-Droid app. "With our software, the soldier can have all those instruments in one device that can be purchased off the shelf. When that soldier returns from war, she can use the same software to protect her family by clocking a speeder near her children’s school and catching the culprit on video.”

In fact a soldier doesn't need all those things: integrated battlefield devices specifically intended for targeting air and artillery strikes have been around for years. They generally have several things that a smartphone lacks - namely night/thermal imaging, proper long-range optics, and laser range finders. Combined with compass and GPS satnav, this means that a soldier can hold a target in his crosshairs, press a button and automatically generate a fairly accurate set of coordinates saying where that target is, which can then be passed over a network of some kind to the guidance systems of one or another type of precision munition: smartbomb, smart artillery shell, or perhaps a Hellfire missile launched from a Predator or Reaper drone aircraft.

A smartphone can't see in the dark: but more importantly, having no laser rangefinder, it can't tell how far away something is with any precision. Wang and his colleagues have attempted work around this by offering a couple of rangefinding methods: in the case of an object of known size, the angle subtended by it in the image can be used to estimate a range - but without any magnifying optics this will offer a fairly shoddy capability at any distance, as anyone who has used manual rangefinders or graticules to estimate ranges will know.

The second option is to take two different snaps from two different locations, allowing the software to work out the target's location using the crossing compass bearings. This will only work for stationary targets and is rather fiddly - and, given the accuracy of typical smartphone magnetometers, will still produce a rather large error ellipse at any distance. Even with an accurate compass it normally takes three bearings to get any real accuracy.

There's also an option to track a moving target on video mode (using the known-size rangefinder option only) and thus generate an estimate of its velocity as well as location.

But in real life, given all these inaccuracies, you would need to call in a pretty major airstrike and carpet a wide area to compensate for your not-very-accurate coordinates: though that has sometimes been an option, in real life. And to be fair to Wang, he does openly say that today's smartphones don't really have the instrumentation for the tasks that his software seeks to carry out.

"Our software is limited by the physical abilities of smartphone hardware," admits the budding war-droid builder, "but the devices are improving rapidly. We anticipate that improvements in GPS accuracy, battery life and camera resolution will allow our software to make even more accurate observations."

Some phones already have proper optics; there are efforts underway to get thermal imaging into handsets; laser rangefinders are already quite small and portable, and could no doubt be improved.

The big weak point is probably the compass, but to some degree that's true in the dedicated military systems too.

So for now Wang and his colleagues' app is a bit previous, though it would no doubt be fun to play with if you could get hold of it (you can't yet, though some details have been published courtesy of the Society of Photo-Optical Instrumentation Engineers here, and there's a press release here).

Given the rather basic nature of the calculations and coding required here, and the fact that not only have the military already done the job in dedicated devices but in some cases commercial app developers too ... you do have to wonder why this is considered valid academic research. ®