nav search
Data Center Software Security Transformation DevOps Business Personal Tech Science Emergent Tech Bootnotes BOFH

Small number of computer-aided rifles could be hacked in contrived scenario

Year-long publicity effort hits bullseye ... in a way

By Lewis Page, 30 Jul 2015

The internet is reeling today at the "news" that a rare make of computer-aided gunsight can under certain circumstances be hacked into, permitting a hacker to interfere with a suitably-equipped rifle's aim.

The gunsight in question is the much-hyped but seldom purchased TrackingPoint kit, a system with a Linux machine at its heart which can be fitted to a range of different rifles.

The TrackingPoint (details on its capabilities are at the end of this article) is mainly a curiosity. People who would be interested in it - experienced long-range marksmen - basically don't need it, and people who need it - those who have seldom or never fired a rifle - typically don't want it. And very few in either group can afford it.

But you wouldn't find that out by reading WiReD:

Put a computer on a sniper rifle, and it can turn the most amateur shooter into a world-class marksman.

The oh-gosh tech mag has partnered up this week with husband and wife security-pro hackers Runa Sandvik and Michael Auger in a publicity push ahead of the Black Hat conference. It seems that the duo have spent a year hacking into the TrackingPoint kit. So what have they done?

Well, the TrackingPoint's Linux heart has Wi-Fi. This is primarily intended to allow video from the gunsight to be exported to other devices, and it is turned off by default. If a TrackingPoint user leaves his Wi-Fi turned off, he can't be hacked.

However, if a TrackingPoint's Wi-Fi is turned on, and left on the default password, anyone who knows that default and is within Wi-Fi range can gain access to the system. This is pretty basic stuff - so basic you would hardly call it hacking, really.

It turns out that one can also, having got onto the gun's local network, get full access up to and including root on the Linux computer, and so do pretty much anything you like to it. It should be noted here that despite the TrackingPoint's trigger feature, this doesn't allow firing a shot: the computer only has the power to prevent firing. If the trigger isn't pulled the gun won't fire, no matter what the hacker may do.

What a hacker can do, though, is tamper with variables in the ballistic calculator so as to change where the bullet will strike:

Sandvik and Auger haven’t figured out why, but they’ve observed that higher ammunition weights aim a shot to the left, while lower or negative values aim it to the right. So on Auger’s next shot, Sandvik’s change of that single number [bullet weight] in the rifle’s software made the bullet fly 2.5-feet to the left, bullseyeing an entirely different target.

You would have thought that having spent a year on this project, the hacker couple would have taken time to look at the TrackingPoint website and bone up on some basic ballistics, but evidently not. However we here on the Reg weapons desk found time to do so this morning, and we would point out that the TrackingPoint system, among other things, compensates for "spin drift" or "gyroscopic drift".

Spin drift is a complex phenomenon and is affected by several factors. However the underlying reason it occurs is that bullets wobble a little as they leave the muzzle of a gun (they suffer "transient yaw", in the trade jargon). This gets damped out by the very strong gyroscopic stabilisation generated by the bullet's spin: but the damping process leaves the bullet stabilised in flight with a small off angle to the right of its line of travel (in the case of normal US-style clockwise rifling twist). This small angling to the right is known as the "yaw of repose" and means that the bullet is pushed slightly right by the air as it flies.

The formula for the angle of yaw of repose is given here. It's a bit fiddly, but what we care about is that the angle of yaw rises with the quantity Ix, which is the bullet's axial moment of inertia. All other things being equal, as a bullet increases in mass ("weight") it has a larger axial moment of inertia, the yaw-of-repose gets bigger and the bullet deviates more to the right ... and the TrackingPoint software corrects aim to the left further and further to compensate.

That's why, if you tell it the slug weighs a lot more than it does, you find your rounds going left of where you expected. Simples.

From the WiReD article it appears that you must actually be in the rifle's local Wi-Fi network to hack it: TrackingPoint's "ShotView" system actually allows video and audio to be streamed across the wider internet, but apparently you can't actually hack the gun from there. You have to be physically close to it and the user has to have enabled Wi-Fi and left it on the default password. Even then, all you can do is alter the direction of aim by quite small angles, brick the gunsight, or prevent the weapon from firing.

We should also remember that total TrackingPoint sales aren't thought to have been much more than a thousand, despite the massive hype around it.

Comment: Maybe not the most glorious bit of 'hacking' ever

So what have we got here?

Basically a very rare system can be hacked in very rare circumstances and this will very, very rarely actually be useful or have any serious consequences for anyone.

This isn't a piece of security research/hacking that was carried out to increase security, then. So why was it carried out?

You would have to say, it looks as though it was carried out pretty much solely for reasons of publicity. The method of feeding stuff to WiReD in advance of public announcement is straight out of the "tech" marketing/PR manual. And it has sort of worked, in that Sandvik and Auger have got their names all over the internet.

But it's possible to suggest that this hasn't exactly covered the pair in glory. It's possible to suggest that a good hacker, like a good programmer, needs to understand not just coding but what the code is doing and how it is doing it.

Sandvik and Auger plainly don't have any firm grasp of what the TrackingPoint system does or how it works, despite apparently spending a year hacking it. They're basically mucking about randomly with something they don't understand.

At least under some definitions of "hacking", that's not actually hacking at all. ®

Bootnotes

The TrackingPoint system includes a laser rangefinder and other sensors, and automatically compensates for atmospheric conditions and range to the target. It also includes image stabilisation and a trigger cutout which won't let the user take a shot until he or she is on target.

What the TrackingPoint can't do anything about is wind, the real bugbear of the long-range marksman. A user can input estimated or measured values for wind, but there's no way to know what the wind is doing all along the bullet's line of flight, so this offers no better capability than the educated guesswork which is - almost always - the only way of compensating for windage.

In fact, then, the TrackingPoint merely automates some tedious but simple calculations which anyone can carry out using tables (or more commonly these days, a phone or tablet) before manually adjusting their sights to get the same benefits. The trigger control compensates for some mistakes that bad shooters tend to make, involving pulling the trigger when the sights are not on the target.

But a TrackingPoint system is heavy, complicated, extremely expensive and has a short battery life, so it has sold poorly - and indeed the company appears to be in serious trouble.

A scope with similar capabilities to the TrackingPoint is included in the US Army's XM-25 Judge Dredd style smartgun, a rather more interesting weapons development, which has been on the verge of introduction to service for some years now.

Experience has shown that some of our US readers may doubt that any Limey journo wiener could know anything of guns, weaponry, tactics etc. So, for the record, your correspondent's canned death-tech CV:

Only significant childhood experience firing a Lee Enfield .303 at the age of five and getting knocked flat on my arse (irresponsible but popular uncle involved). Member of the pistol club at university, mostly shooting .22 target weapons but heavier pistols on occasion (a while ago obviously, that).

During early Territorial involvement gained some familiarity with the old British 7.62mm SLR (aka FN FAL for non Brits), 7.62mm Light Machine Gun (Bren gun but shooting NATO rounds), 9mm Sterling submachinegun, all now long retired.

Eleven year Royal Navy service career, for most of which I had to stay current on the Browning 9mm pistol and 5.56mm SA80 rifle/"Individual Weapon" (for most of that time, the widely reviled L85A1 version rather than the new good A2 from Germany). Commando qualified, and as such also thoroughly familiar with the SA80 L86A1 "Light Support Weapon" (nowadays regarded as a marksman weapon rather than squad-auto, as the L86A2); also handled and fired 7.62mm general-purpose machine gun (FN MAG for non-Brits). Some idea how to use that latter in the sustained fire mode with dial sight.

Other things I have fired but do not know well at all: 30mm Oerlikon cannon, Sig-Sauer 9mm pistol, .45 Colt pistol, H&K MP5 submachinegun.

Also acquainted with many other kinds of weaponry, almost all other kinds in fact, but mostly from the point of view of the Explosive Ordnance Disposal operator rather than the user.

Have also fired 12-bore shotguns on a few occasions and was once even present at a grouse shoot.

The Register - Independent news and views for the tech community. Part of Situation Publishing