Terrorist robots dissected - anatomy of a scare
DIY cruise missiles - not as easy as you think
Analysis There's a fresh flurry of robot-terrorist headlines on the wires, following an autonomous-weapons conference held on Wednesday at a military thinktank in London.
The assembled military and academic bigwigs gathered in Whitehall were mainly thinking about the ethics of killer robot use by Western forces, but there was only one thing the media were interested in - terrorist robots, probably flying ones.
AP quoted  Rear Admiral Chris Parry of the Royal Navy as saying: "Sooner or later we're going to see a Cessna programmed to fly into a building," and suggesting that remote-controlled aircraft - perhaps converted from toys - were "ideal [terror] weapons... They are cheap ... and they're difficult to detect — about as difficult to detect as a blackbird".
Admiral Parry is already mildly famous for predicting a Roman-empire style collapse of Western European civilisation, caused by Goth-like waves of migrants who fail to assimilate and remain connected to their home cultures by the internet and cheap flights. (We still think he was referring to Australians, Kiwis, South Africans and so on.)
The good admiral also presided over a MoD crystal-ball report  last year, which seemed to predict a global middleclass rebellion against a power structure dominated by the super-rich. The admiral and his fellow report drafters seemed to suggest that this rebellion might make use of flashmob riots, electropulse blasters, or even brain chipping.
They also hinted that: "A cheap, simple-to-make and easy-to-use weapon might be invented that is effective against a wide range of targets."
Now Admiral Parry reckons he's found that weapon. He's not alone, in fact; various people have been warning about the possibility of terrorist or other bad hats building themselves a DIY cruise missile for some time now.
In fact, the thinktank discussion assumed as much  (Word doc) to begin with:
"With small or mini Unmanned Air Vehicles (UAVs) available for as little as USD$1000, the advantages of autonomous vehicles could easily be harnessed by militias, criminals and terrorist groups..."
Indeed, robopocalypse prof Noel Sharkey reckons  he could do better, and build you a flying deathbot for just £250.
Is it goer? Well, sort of. Civil GPS is now cheap and easy, so your autonomous system can certainly guide itself to a known target location - let's say No 10 Downing Street. GPS is completely passive, which is nice as it means our homemade cruise missile won't be entirely simple to detect or meddle with electronically.
However, a £250 to £500 UAV  - while it can surely fly a GPS waypoint track - can't lift any significant warhead. You could use it for spying or something, though anyone with a shotgun will be able to shoot it down if you get too close. This is not a killer robot.
But let's assume a bit more funding, just to be sporting.
You can buy a small, petrol-powered robochopper  which might cost under $60k, carries up to 50lb of payload and could fit into the back of a van or small lorry all ready to fly. Pull up somewhere quiet the other side of St James' Park, launch the bird and drive away.
It has controllers and software capable of automatically flying a 3D waypoint route, apparently  good enough to stay inside an area no more than six metres wide and one metre up and down.
So the robochopper zooms along gaily without further input from you just above the rooftops and trees, crosses Horse Guards and comes in where you choose - roof, window, maybe doorway of Number 10, accurate to maybe +/- 3 metres. Fifty pounds of explosive device goes off. Job done.
Maybe not, though. You've definitely got your headlines but you're fairly unlikely to have bagged your prime minister, unless you could predict in advance - almost to the second - just when he'd be between the door and the armoured car (even his press secretary probably can't predict it that closely). Even then, a bit of wind, an alert close-protection man, and you're out of luck. He'll be into the car or the building before the explosion.
One can, of course, destroy or partly destroy buildings and armoured limos with less than 50lb of explosives - the Brighton hotel bomb was smaller than that - but the stuff has to be carefully placed and prepared, probably inside the target. If all you can manage is crashlanding the bomb within an area six metres across you need to be thinking in terms of hundreds of pounds warhead weight for these targets.
The new, US "miniaturised" GBU-39 smartbomb - ultra-accurate, designed to get a kill using the minimum necessary amount of bang - is in the 250lb class. Most of this is actually the steel penetration case, which would let it punch inside No 10 or the PM's limo before detonating its 50lb charge. But the robochopper can't hit hard or fast enough to get this kind of result, even if it could lift the weight.
Probably the best plan would be to make a 50lb large-diameter shaped charge of the sort now being used in roadside "superbombs" in Iraq (it isn't too hard) and hang that under your robochopper. Fly the chopper at the PM's vehicle. To have any serious chance of success you will need to manually pilot the weapon in; GPS will quite likely miss altogether even if the vehicle is obliging enough to remain stationary, and with an explosively formed slug you need to be fairly accurate.
A $50k-60k robocopter can be handled remotely using a live vid feed, but now you've got serious problems. Your drone is no longer electronically stealthy - it can be detected a damn sight more easily than a blackbird. It can also be meddled with, and it will be. Even ordinary military patrols these days deploy sensors and jammers which can warn of an elctronic threat in advance and perhaps cancel it out.
The RF signals/EW battle is one that terrorists will really struggle to win. Lose badly enough and you won't just fail - you'll be traced to your remote piloting location and the next thing you know you'll be hip-deep in SAS men. To be honest, roadside emplacement would probably work better - perhaps that's why people tend to do it that way.
Even a nice simple GPS autonomous drone - if you could somehow make it strike hard enough at a reasonable price and size - can be jammed or spoofed without difficulty. It isn't hard to blot out or degrade the genuine civ-GPS signal within a smallish area, and the level of inconvenience to those nearby isn't all that great. It's not like drowning out hospital beepers or something. And the nasty old western forces aren't bothered at all, because they can use the encrypted military GPS signal.
Frankly, if GPS drones ever seem likely to become a threat, you can expect cheap simple GPS jammers at every target location. This sort of thing is one reason why a Tomahawk cruise missile costs hundreds of thousands of dollars, not $1k. (The other big factor is the need to lug a big warhead and lots of fuel.)
GPS deathbots aren't likely, though. Not soon. Terrorist cells which can make or obtain 50lb of reliable explosives, build them into a $60k robo-aircraft without blowing themselves up, and then operationally deploy the system by lorry without being betrayed, are quite thin on the ground.
If this was the heyday of the Provisional IRA, you might want to think about those GPS and RF-video jammers; but it isn't. The new kids on the block aren't in the PIRA league. They operate in UK organisations typically 10 strong or less - probably because they don't have solid community support - and it's a big day for them if they can make risky-but-functional TATP backpack bombs. Mostly they aren't even that good - indeed, they're often almost comically inept .
Even if PIRA were back, so what? Accept a little bit of GPS jamming around obvious targets, and get on with life.
Once again, it seems that the main thing we have to fear is fear itself. ®