From aerial torpedoes to RoboCarsPublication date: 21 October 2009
Originally published 2009 in Atomic: Maximum Power Computing
Last modified 03-Dec-2011.
Combat robots were science fiction until, suddenly, they were on the evening news.
We haven't quite made it to the Terminator/Cylon/Tachikoma stage yet, of course. Drones are not yet turning themselves into humans and interrogating people. It's more pilotless planes (Unmanned Aerial Vehicles, or UAVs), and pilotless planes with guns (Unmanned Combat Aerial Vehicles, or UCAVs).
The new wave of UAVs didn't initially get a lot of media coverage, partly because nobody was crazy about letting the world see pictures of their secret recon drones, and partly because UAVs just aren't very... glorious. Ex-flyboy Air Force bosses have never been crazy about letting some geek with a lot of monitors and a retail computer joystick do what's traditionally been done by strapping young men dripping with testosterone. They can just about tolerate recon drones and bomber-imitating decoys, but anything that resembles an unpiloted fighter upsets them.
Autonomous machines have actually been sneaking up on us for rather a while, and we're making serious inroads on the much more difficult problem of autonomous ground vehicles, too. Those were completely out of the question in the early days of UAVs, which were, believe it or not, back in World War I.
UAVs then, steered by gyrostabilised autopilots, were called "aerial torpedoes". (The same term later meant a normal naval torpedo dropped from an aircraft.) These early UAVs didn't make a great impact on the military world, or on the enemy for that matter. But they gave rise to all of today's "smart" weapons.
The WWI aerial torpedoes were, in fact, primitive cruise missiles - and a modern cruise missile can independently fly a complex course of more than 2,000 kilometres at little more than treetop height. That sounds like a pretty impressive UAV to me, even though it's not the kind that needs landing gear.
The nice thing about the sky is that there's not much to run into up there, so a cruise missile that has very close to no brains at all, like the WWII V-1 flying bombs, can be effective. Unmanned land vehicles have to be able to tell the difference between a shadow, a bush and a rock, though, and you ain't gonna do that with clockwork and gyroscopes.
If you can solve these problems, though, then you're making devices that can assist, or even replace, ground soldiers. Classically, this has meant robot tank projects - but now all sorts of smaller oddball things are showing up, from iRobot's brainless but useful PackBots, though the larger, general-purpose Gladiator TUGV and semi-autonomous bomb disposal robots, to the creepy bouncy pantomime horse that is the "BigDog". University teams have produced all sorts of inexpensive UGVs and UAVs, and small remote-controlled UAVs are becoming useful tools for infantry too. That tech's getting cheap enough that a hobby is breaking out.
(Many people have observed that, today, it's really not very difficult to make your own small-payload, short-range cruise missile. One large model aircraft, one microcontroller, one GPS, one pipe-bomb warhead, add software, mix thoroughly. The fact that such weapons have not been raining down on innocent civilians across the Western world is just one of many indicators that domestic terrorism is not at all as serious a problem as some people would like you to believe.)
True self-driving Unmanned Ground Vehicles still seemed pretty hopeless when the world first got to see a lot of them in one place, at the 2004 DARPA Grand Challenge. That race had 15 starters, but zero finishers. Not everybody even made it past the audience stands at the starting line. (It reminded me of that bit in Robocop 2 with the less-than-successful Robocop-2 prototypes.)
Here's a whole documentary about it:
The very next year, though, five of 23 starters made it to the end, and only one didn't get as far as the best competitor from the year before.
The next challenge was held in 2007, in an urban environment including other traffic, instead of the arid back-country of the previous races. The final race was roughly as difficult as any ten really annoying Grand Theft Auto missions put together, but it still had six finishers from eleven starters.
On the military side, "smart" vehicles and munitions are yet another force multiplier. Precision weapons do make it easier to kill enemies hiding among civilians without as much "collateral damage", and robot planes and tanks do let you keep your own soldiers out of harm's way. But the result of new force multipliers never actually seems to be "less war". The easier technology makes it to achieve a military objective, the more likely someone is to try it.
The civilian version of a UGV is the driverless car, though. And, as I've written before, the world's yearly road toll is in the order of a million deaths. People are not good at driving cars. Robot cars can't come soon enough, if you ask me.
The only thing that depresses me about the robot-car future is the substantial legal, not technological, obstacles in their way. A big part of that is making it legal to actually put robot cars on the road, but there's something else waiting when they get there.
If, you see, someone drives a Camry over you tomorrow, you (or your next of kin) are unlikely to sue Toyota. But even if the motor-accident rate for driverless cars is one per cent of the regular-car rate, people who get hit by a RoboCamry - and, heck, maybe also the people who're sitting in it when it hits someone - will sue Toyota.
This situation is not unprecedented; airbags had the same problem. If, for instance, you're looking for something in the glovebox when the passenger-side airbag goes off, that airbag will try pretty seriously to kill you. Lawsuits have been filed. But airbags still managed to become standard equipment.
I've got a better solution, though.
Military customers seem to be pretty happy with anything that can launch missiles, even if the rest of the vehicle doesn't work very well. I'm sure ordinary consumers feel the same way. So the New 2015 RoboCamry just has to come with Three Free Hellfires!
I admit that this might cut into the road-toll improvement a little. But you can't say it wouldn't make driving exciting again.