Robo-warship sub hunter: Free DARPA crowdsauce game No doubt regular readers will recall the US military's cunning plan to develop unmanned submarine-hunting robotic frigates - warships which would prowl the oceans like automated Mary Celestes, remorselessly tracking enemy submarines regardless of how their pale, sweaty, malodorous captains* might twist and turn. Screenshot …
One can safely assume that ridding the oceans (or indeed the planet) of Branson would be met with general rejoicing. I mean, come on, if you are a hitherto unknown trench-dwelling creature of the endless night, who would you least like to be discovered by...
Algorithms will need to include an annoying beard (AB) and grating self-importance but ultimate irrelevance (GSIBUI) detection filter. To avoid needless slaughter of marine researchers, single seater deep exploration subs are ONLY to be attacked IF AB = Branson AND GSIBUI > Clegg
"putting up snorkel or 'snort' masts (to run diesels) is suicide as airborne radar can sweep huge stretches of ocean"
Right, so how about creating a bunch of mini-torpedos which you set to quietly motor off in various directions and, after a set amount of time, they stick up a mast with a radar reflector on the top and suddenly you've got snorkels everywhere! (Oh, and add in a sound generator to confuse the Sonarman Jones's of this world ;-) )
As for "Responding manned units starting from the "flaming datum" of the destroyed ACTUV would soon find and destroy the sub as it crept away - all the sooner if it surfaced or put up masts in order to flee more quickly."
Right, another mini-torpedo, but this one's job is, for instance, to latch onto the ACTUV with a magnetic grapple and then steer it off in another direction or just apply reverse thrust to keep it in place for a while then, after a set period of time, explode.
For every "cunning plan" that someone comes up with, there's an equally cunning counter :-)
The Royal Navy used fake periscopes in the Mediterranean in WWII, just sticks painted grey and weighted at one end so they floated upright. These were dropped, normally by submarines, at choke points where enemy vessels commonly passed. They were not that successful as a periscope is hard to spot at the best of times and as they were immobile they produced no telltale feather or wake. I don't think the axis powers developed any cunning countermeasures.
"In any case, an ACTUV chasing a sub could easily be said to be engaged in "underwater operations" and thus claim the status of a vessel "restricted in her ability to manoeuvre" under the Regulations - other ships would be obliged to get out of her way."
This would raise the potential of maritime trolling of the internet sort rather than the (original) fishing sort - one or better yet two or three submarines could thereby disrupt a port's relatively narrow shipping lanes by 'towing' their pursuers through those lanes in semirandom zigzagish paths.
That being only one of several downsides, I (as a former surface line naval officer) agree that complete 100% autonomy just doesn't seem to be a worthwhile goal for these platforms.
It really is long overdue that the current 'western powers' stopped basing their defence on oil burning ships, submarines and tanks.
A diesel sub needs air to flee any distance effectively, but a nuclear sub doesn't. On those grounds alone the Royal Navy should by now have swapped all its subs to nuclear power. An aircraft carrier needs lots of oil to shift its massive bulk through the sea - so we should be building the QE class to use nuclear as well - rather like the big American ships. Yes, in a war there is a chance that the sub/ship will be sunk with the long term leak of nuclear waste, but the alternative is that the UK finds itself with its one serviceable aircraft carrier drifting out of oil and the subs uselessly tied up being refilled so has to turn to nuclear missiles to defend itself - far more radiation than if we built the nuclear ships/subs.
Given that most diesel subs travel close to their own territorial waters, if the sub captain wanted to shake his pursuer, he could easily sail close to home, leaving roboship miles away, then go down the coastline for a few miles and then back out to sea again.
If roboship is dumb enough to follow the sub into territorial waters, then blow it out of the water.
The USA knowing where you are isn't suicide - they're not allowed to attack you without a good reason.
However, the USA knowing where you are *and* knowing that you blew one of their vessels out of the water is suicide, because now they can retaliate.
During the cold war there were loads of pairs of Russian and US subs following each other. They knew where each other was most of the time, but nobody was going to start anything because the other side would finish it. It's basically MAD on a smaller scale.
The point of the ACTUV is that you know where the sub is, and it doesn't matter if you go active and the sub knows you're there, you just want to track it.
However, unless the ACTUV is armed and able to defend itself, then its relatively simple for the sub to take it out, and so it would be redundant. To overcome that, it needs weapons, radar, sonar and a firing control computer. As a fully autonomous weapon fire and control system would keep safety engineers busy for decades trying to clear it for use (to say nothing of the treaty that prohibits autonomous fire without human intervention to prevent a 3rd world war being between robots without consequences), this means that weapon fire would need realistically need to be remotely operated (a la predator), which also means comms, which means it can be jammed by a buoy from the sub - even if the ACTUV has broadband, LPI, LPD radio. In which case, they might as well drive the thing remotely too, making the autonomous part redundant.
DARPA = where all the nutters go when they're tired of the real world.
So in order to develop an algorthym for tracking subs they've programmed a simulation of some subs which try and get away. They're then going to base the tracking method on how players follow these AI subs.
Surely the outcome of this is only going to be as good as the AI used for the sub's evasion tactics?
It'd work better if they'd either put humans in charge of both vehicles or spent the effort put into the sub AI as a first iteration of the tracking AI and then seen how human players tried to evade it.
The presumption here is that destroying the perusing robotic craft would result in a swift destruction by a surface fleet but that would seem like an over-reaction. Would any nation really kill dozens or hundreds of submariners in retaliation for the loss of a piece of hardware? Would they risk a presumably major* war for such a transgression?
Without the unrealistic threat of destruction, what would stop a sub captain from destroying the craft, if only for the target practice or from a sense of annoyance? And wouldn't the same fate await all it's AI-piloted replacements whenever they showed up on station above him?
* Any nation currently able to put a submarine to sea will probably have a well-equipped and suitably modern military. Even more so if their submarines were considered worth following in the first place - e.g. Nuclear launch platforms.
Even in the case of a submarine used for smuggling, does the destruction of a robotic craft involving no loss or threat to life justify killing all those on board the sub? I can't see that it does. So all ways around, there wouldn't seem to be any sense to the scenario described by Lewis where the sub-hunting robot remains permanently in place above the submarine and destruction of the robot would carry a death penalty.
More to the point, authorities normally wish to intercept smuggling operations to board and make arrests ASAP. When they do intercept them they usually aren't concerned about their frigates being torpedoed. I might imagine these craft being used to find smugglers subs, saving on manpower, but not following them around for months ... or even days.
"...does the destruction of a robotic craft involving no loss or threat to life justify killing all those on board...?"
-
Depends on what particular Sole Remaining Superpower owns it now, doesn't it? Ask Mistuh Tony Bliar, ol' 'Known Unknowns' Rummy, any of the Cheney+Bushite Vulcan Gang and/or the top-tier Red-Shield Financier for that bunch etc ad naus (if y'reckon any of 'em'd ever deign to talk to you.)
-
Means, Opportunity, Motive. Really, don't go bothering any of the above-mentioned Rich, Powerful and Very Shiny warhawkers. ( [Pa-TOO! ]) Just ask yer MOM instead... And that is all. 0{;-|o<
Perhaps the point of their game is not as simple as they admit? By encouraging players to operate their machines, they're also studying ways to counter those tactics. AI systems are still very complicated and known to do some very strange things sometimes. Is suspect they're trying to refine the AI and also prepare for the day Russia, China, India or somebody else starts developing machines like this.
If this thing is going to sit on top of your sub, pinging away with active sonar, wouldn't it be rather trivial to send out a scuba diver with a mine to attach to it? You could then program this to detonate when the frigate got tired of tracking you and got, say, 100 miles away.
I admit it would take balls of steel to carry on the rest of your tour with a live mine above your heads, but it sounds like it could be a plan.
Downloaded. Tried it. I have no problems giving the data to DARPA. Clever idea collecting data from players. Low cost R&D ! The problem is - it's no fun. This is from some one that played 688 Attack Sub for days and loved that game. This has a 57 page manual written for nuclear physicists. Bottom line is that it's like watching paint dry. Does not look like you ever get to kill anything <yawn>. Nice idea DARPA. Try again.
Think about it - the frigate has to manoeuvre to avoid surface shipping. Sub-driver surfaces, launches a couple of RIBs. Submerges again and moves off in a direction that leaves the RIBs directly between it and the frigate. When the frigate turns aside to pass them, the RIBs move to block. Frigate turns aside more to pass them, they block more. Sub gets away.
This post has been deleted by its author
When being trailed by one of these robo-subs send up your frogman with a bag of King Edwards and get him to jam them in it's exhaust. Couple of minutes later robo-sub will cough to a halt, simples.
A large thermal exhaust port may require the use of seasonal squash or potentially even a marrow.
Anyone here ever play CRobots? This story triggered recall of my college days (late '90's) where our computer science class competed to come up with the best algorithm for a simulated autonomous robot that could move, scan a certain direction, and fire a projectile. I think the later versions also added IFF and communication between friendly robots. Coding an autonomous swarm would be pretty fun... :)
Diesel submerines don't sit on the surface with snorkel pocking out 24/7, the engine will be used to charge the giant batteries every now and then i.e when safe to do so. If a threat is detected the sub turns off eangines and goes silent, surely an unmanned sub would make even less noise as it doesn't need to filter air, cook dinner etc...
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
