Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
The US Navy and Your Autonomous Vehicle
Damage to the USS John S. McCain
The United States Pacific Fleet seems to be having a run of bad luck. Or is it?
There have been five major accidents in the last 12 months.
- August 18, 2016: The USS Louisiana (SSBN 743, Ohio Class Nuclear Submarine) collided with a USN support ship, the Eagleview (T-AGSE-3) in the Strait of Juan de Fuca, the waterway between Washington State and British Columbia.
- January 31: USS Antietam (CG-54, Ticonderoga-class guided missile cruiser) ran aground in Tokyo Bay.
- May 9: USS Lake Champlain (CG-57) collided with a South Korean fishing vessel.
- June 17: USS Fitzgerald (DDG-62, Arleigh Burke-class destroyer) collided with the Philippine freighter ACX Crystal.
- August 21: USS John S. McCain (DDG-56) collided with the oil tanker Alnic MC.
How does the world’s most technologically advanced navy keep hitting things? The answer may be found in the question itself. Two separate Navy officials told CNN that the McCain experienced a sudden loss of steering control right before the accident, only for it to reappear just as suddenly afterwards.
The McClatchy news service reports that on June 22 in the eastern Black Sea someone highjacked the GPS capabilities of some 20 vessels. Their navigation systems, all of which were operating fine, suddenly placed them 20 miles inland near an airport. This is the first reported instance of widespread “GPS spoofing.”
The Navy acknowledges that they may have been hacked. Chief of Naval Operations, Adm. John Richardson said there are no indications of a cyber intrusion in any of these instances at the present time but the Navy remains open to the possibility. It’s not a new thought, either. In 2006 the Academy at Annapolis ceased teaching celestial navigation but the sextant was returned to the hands of the Middies in 2016 although the Navy denied at the time that they were worried about hacks.
As Google and vehicle manufactures begin a serious push for autonomous cars and trucks think of the chaos and destruction a terrorist or foreign government could inflict in one day if enough driverless vehicles were on the road at any given time. Somebody needs to be in control. Right now, it’s not clear that is us.
Published in Technology
Thanks for writing about this.
I was wondering why we never hear more bout who controls the fishing boat, or container ship , or whatever other vessel rams our battleships in onetragic accident after another.
I don’t believe any of these incidents are just chance encounters.
I don’t think it is very comforting at all to realize that we have met the enemy, and he is us. Except that it may be easier to correct our own errors.
You are right to conclude that this is a systemic issue, and not a coincidence. I suspect a lack of training and discipline, which are the results of decisions made at a high level over a period of years — 20 years, or more. We have held ship COs accountable — and properly so. But we have not held the senior Navy leadership accountable for their performance. The rot runs deep.
What I’ve never gotten an answer to is the question of whether a self driving car must of necessity be hackable. In other words, whether, in order to be autonomous, it must be hooked up to the internet or otherwise designed in a way that it can be taken over remotely. It doesn’t seem that the one would necessitate the other. But I’ve never heard of any of these new fangled cars not having a hackable interface of some sort. In my opinion if they can’t make a self driving car that is unhackable…meaning that there’s nothing to hack…then they shouldn’t make self driving cars. (And if Navy ships are hackable, whoever is responsible for that should be exiled to Antarctica.)
The GPS systems operate via external communications, so by necessity requires an external interface rather than a closed system.
So if there’s a countrywide internet outage, every autonomous car on the road crashes or stops?
Not necessarily internet. Internet requires Internet service providers and access to servers in various locations around the world. I think GPS uses triangulation by pinging off of satellites and telecommunication towers to determine your global location that is then superimposed on a downloaded map (that downloads and updates when you have internet connection).
When you drive, you don’t always have internet connection, but the gps system continues to work.
Actually I have a decade old car with a built in GPS system. It’s not even theoretically hackable I’m virtually certain. There’s nothing to hack. But it does communicate w GPS satellites and so it seems a self driving car could be built to do that and guide itself accordingly without any other interface connecting to the controls.
If it requires external communication, it can be interfered with.
Come to think of it GPS probably isn’t how self driving cars maneuver. GPS couldn’t be accurate enough to tell a car when to brake if the car ahead stops or slows for example. The on board censors must control most of the driving. So it seems possible to make such a car a closed system with a built in map in its software.
But not the controls to the car. My car’s controls aren’t connected to the GPS. Only the map screen which shows me where I am is, and I control the car.
GPS does not have an external interface. It is a sensor that senses the timing signals from the satellites. In this sense a standard GPS receiver cannot be “hacked” no matter how good the computer hacker is unless they have physical access to it. It doesn’t accept commands or connect to anything like the internet.
Some more modern GPS receivers do connect to the internet to download corrections to the pre-programmed satellite orbital parameters, but this only improves the accuracy. Without these corrections GPS is accurate only to a meter or so – still plenty good enough to navigate a boat or ship. The GPS navigation system on a Navy ship should not be connected to the internet. A standard GPS receiver without the corrections is plenty good.
There are other ways that someone could mess with GPS, but these are much more sophisticated and involve using broadcast radios (and the one in your laptop or cell-phone is not good enough – yet). With the proper software defined radio you could jam GPS by broadcasting at the GPS radio frequencies, or even more sophisticatedly spoof GPS by broadcasting GPS like signals that the receiver mistakenly takes as a satellite and so causes it to calculate the wrong position.
The fact is that boaters too often overly trust their GPS system, but that shouldn’t be happening in the Navy.
As far as self driving cars. The systems in use today use GPS to tell them their general position (i.e. sometimes down to which lane the car is in) and when you need to make left turns, etc. But they use camera systems and radar systems to keep the car in its lane and to navigate the actual turn and to avoid obstacles.
I just had an interesting job of the future – remote car driver. Basically, no matter how smart the autonomous driving system it will encounter situations, such as a truck parked in it’s intended path, that it won’t know what to do. So in those cases when it doesn’t have a passenger in the vehicle, it will just stop safely and ring up a remote driver that will assess the situation and make a decision on what to do. Kind of like a drone pilot.
This is what I was thinking of. I wouldn’t expect GPS “hacking” to be done by an amateur hobbyist mischief maker.
It would use GPS for big picture information and sensors around your car for navigating obstacles. Your car’s internal “pedometer” (what is it called, that ticks with each mile?) would do some work as well.
The Russians may be doing something along the lines of spoofing GPS systems.
The destroyer should still have been running its radar. Messing with the GPS wouldn’t have changed that. There has been the suggestion that there was a problem with steering. We don’t know enough yet.
So, the way the Nork ShortStack’s been acting lately, we could speculate out as far as, say, this weekend?
Yeah; it’s the location more than anything that has me worried about that one. GPS spoofing is much like shouting over someone’s conversation. It works better when you’re near by. The black sea is physically close enough to Russia that they might be up to something. If this happened in the Bay of Biscay I wouldn’t be half so suspicious.
None of these civilian systems are truly closed, are they? They have to be accessible to software updates and information on traffic congestion. If you could spoof the system into believing it was somewhere else an 18-wheeler that decided to take a non-existing exit could create a lot of carnage, no?
Question to our Navy vets and technophiles:
If you seriously wanted to jam or mess with nav systems would a ship mounted transmitter do the trick? And without a pretty impressive stick how close would you have to get?
I don’t have anything technical to add, but when we drove next to a Tesla whose driver was sleeping (!) in rush hour Honolulu traffic, we were both awed and envious.
Yes, I know Mr. Goldfinger, enemy action!
Not sure about these sources. I don’t think they really capture the right facts to really explain the situation.
You have it right JudgeMental, as far as I know, it wasn’t anything to do with pilot error, it was a interesting startup sequence practice stemming from taking shortcuts (not using two independent safety measures as required by the manual) and easily failing equipment (dislodged disarming pins) plus a electrical surge anomaly that few should have predicted.
Also, from what I understand, subs really do have a really hard time with navigation. There is no way to look at the surface in a sub, you have to get to periscope depth first, the pop up depending only on sonar and whatnot. Also, subs are actually not very maneuverable on the surface. They are meant to be underwater, not above. So, subs colliding with tenders or other surface ships, whether navy or private, isn’t much of a surprise. Also, running aground seems like a similarly hard to avoid thing.
So, when I hear that we have had about 1 collision a year, it doesn’t really sound that bad considering the number of subs we have.
This is, of course, not to say we shouldn’t work harder on avoidance.
And, while I am not too concerned with the subs, I am worried about the surface ships. I will have to read the full report since it does sound like there where many issues with the Fitzgerald.
I didn’t get this the first time around. Kudos, sir.
I don’t think you’d have enough situational awareness to take control.
When I was in the Navy in the mid 90’s on frigate there were five ways for the ship’s navigator to get a position fix. As far as I know only Loran-C has gone away. The surface search radar was always running, and it had a display on the bridge the OOD and JOOD could look at, and a duplicate in CIC that was manned and the watchstander switched out every hour. The radar operator, the external lookouts, and a bosun on the bridge were all on the same voice network so that any contacts by eye or radar would be reported.
As far as the steering, there was a control on the bridge that operated the hydraulic ram that moved the rudder. There was also a redundant control in the steering gear room that could be used to steer. If the hydraulics failed there was a a ratchet mechanism that could move the rudder, and finally if you were desperate enough, there was a huge wrench that you could put on the rudder top and several guys would steer the ship with a tiller. If even all that failed, you could shut down the main engine, and extend two motor pods (Auxiliary Propulsion Units, APUs) the could independently swivel 360 degrees.On those alone the ship can go four knots and turn any direction.
So I have a hard time buying that there wasn’t grievous human error going on here.
Hydraulics are notably resistant to hacking. They don’t even network for heaven’s sake.
This also struck me as strange. We know that China wants to displace the US Navy from the Pacific as the primary naval power. It was surprising to me that all of these situations involved the Seventh Fleet headquartered at Yokosuka.
And not only in the Navy, by all accounts. What can be done in the lifetime of a single administration to at least chip away at this rot?
And thus the training issue. Maybe.
Are you sure? An SSBN is a strategic asset. Wouldn’t it be assigned a strategic mission under the Navy version of SAC?