An airline that has made sure its airplanes won't disappear | Hacker News Reader

An airline that has made sure its airplanes won't disappear

washingtonpost.com

61 pointsLibertatea60 comments

I deploy satellite-comm GPS tracking equipment on machinery in the ski industry.

What I don't get is why airlines insist on bundling emergency telematics along with position tracking. It massively complicates what otherwise is a pretty simple problem. It feels like that description of the Bradley Fighting Vehicle out of the Pentagon Wars movie.

With a box the size of a Big Mac, at <$1k per unit and about $100/month in data costs, you could deploy an off-the-shelf industrial grade tracking unit that would transmit location/speed/altitude every 10 seconds whenever the unit is powered. All it would need would be a power supply. This would neatly solve the "where the heck is the plane" problem, globally, for a fairly trivial cost.

I realize that finding a spot on the plane to mount it, verifying aerodynamics, figuring out how to wire it in, testing, doing the work across a fleet and managing all that data all end up costing a lot more than the equipment, but given the costs being quoted for some solutions in the articles that have been coming up, it really seems like the airlines could have a big PR win by paying for it with petty cash out of their marketing/advertising budget.

I'd like to talk with you. Could you hit me at my email in my profile?

Email sent.

Could you describe a bit which type of satellite communication you use for that the ski use-case, and what is available as off-the-shelf equipment ?

Sure. I run a small business (http://skitrails.info/) offering real-time GPS tracking of grooming equipment at ski resorts. The resorts use the data for operations management, and to prepare reports for public consumption (i.e. what's groomed).

I'm mostly using cellular/mobile vehicle trackers (of which there are hundreds of vendors), but for regions where there's no cell/mobile coverage, we fall back to satellite options. With lots of optimization, it's only slightly more expensive for our use-case.

Specifically, we deploy these units:

http://www.skywave.com/en/our-technology/satellite-cellular-...

which use the Inmarsat constellation. The specs probably aren't quite up to being attached directly to an airplane, but with the right housing and a heater, they're probably not far off.

The trick for finding robust off-the-shelf gear is to look for industries that operate in remote regions, like oil and gas, and maritime. They also have fairly harsh operating conditions, so equipment sold there is usually fairly tough.

Prices have really come down over the last 5 years.

I wonder if there's a fear of good but partial solutions. Maybe they think "we couldn't add tracking because it's too costly" is better than "we added tracking, but it didn't transmit the one parameter that would be crucial to finding the plane we just lost." To be clear, I disagree, but I could see being paranoid of bad PR that way.

I don't like the tone of this article. Articles like this were written after each and every single crash, which eventually led to sharper requirements, more certification, added security gizmos and thingies and more safe-ropes in every single corner of the aircraft. Most of that was reasonable, but care must be taken when making systems more complicated and static by taking what seems to be the easy route of "just adding a fix".

The reason why most of these systems seem anachronistic is not due to the high costs and financial pressure on the airline companies. Most systems seem so anachronistic because they have to be incredibly reliable and robust. Flight recorders are built to withstand about any disaster you can throw at them, which would be quite hard to do achieve with a modern cloud-enabled micro-electronic device. Even if we threw billions at the problem, it would most likely take five years to build a mass-deployable remote aviation telemetry system of the required quality level, just because making hardware is hard, and a half-assed solution helps nobody. It's particularly unlikely that a niche solution like the one suggested in the article works for everybody.

Don't get me wrong -- added safety is always nice. But let's take a step back for the moment. Let us not just put another whole new box into every aircraft after every crash. Instead, let's try and make the smallest possible adjustment to what is a well-proven, existing system already.

There have been reports after the MH370 loss that turbine telemetry data had been transmitted even after all other systems had failed. An small adjustment would be to lower the interval in which this signal was transmitted, down to say 30 seconds, provided the comm satellites are prepared to handle this. I'm not an expert on avionics, and I'm not claiming that this is the egg of Columbus, but surely there has to be a nice, small, lightweight and elegant solution, which incidentally is cheap on top of this.

Agreed. People forget that in addition to the nominal cost of a new tracking system ($120k PER PLANE in the article, not including control center cost), just the weight alone costs about $3k/year for each 1kg added to the plane. Add the yearly communication service, maintainance, depreciation costs to this, and we're soon gonna be paying $3k/plane ticket.

In contrast, since MH370, Inmarsat has changed the time period for handshakes with their terminals from one hour to 15 minutes. Cost: $0.

> $3k/year for each 1kg added to the plane

sorry, offtopic, but this is an interesting number. Do you have a source for more on this? Most of the search hits I can find are discussing economics of spaceflight.

Usually you find these kinds of figures in articles about 3D printing aircraft parts. For example, "By reducing an airplane’s weight by just one kilogram, an airline can save $1,300 in fuel costs per year, according to the [2011] IBISWorld report." [1] The 2014 report can be purchased here [2]. Of course, these numbers will depend on the kind of plane and its utilization, as well as fuel costs.

[1] http://www.businessweek.com/technology/3d-printing-coming-to...

[2] https://www.ibisworld.com/industry/3d-printer-manufacturing....

On the other hand, airlines make tens of billions in extra revenue from those little fees to make it just a little bit less like a cattle trough.

I don't understand why such a difference between what you are saying and a portable SPOT device [1] which is pretty much standard for extreme sports nowadays. A SPOT device is only 114g, tracks every 2.5 minutes, and runs on batteries for days. $170 for the device and $200 for the service per year.

Actually if passengers start carrying these devices (while sitting next to the window) that's already a solution.

[1] http://www.findmespot.com/en/index.php?cid=100

There is no doubt that it would be costly to add real-time tracking and telemetry. However planes going missing is also very expensive.

I have to wonder if airlines would be more accepting of the costs of monitoring equipment if they paid for search and rescue themselves instead of relying on governments to cover the costs. A classic case of privatizing profits and nationalizing costs.

Wrong, commercial passenger planes going missing are much less expensive. Commercial plane disappearance is close to a 'black swan' event (1/10years) [1]. Tyler Durden this:

* search and rescue for this rare event: $50M/10year = $5M/year

* cost of tracking system: $20k/year amortized cost per system * 5k aircrafts = $100M/year

The airline industry as a whole has made a cumulative loss during its 100-year history. Airlines are already nationalized one way or another, like it or not. [2]

[1] https://en.wikipedia.org/wiki/List_of_aerial_disappearances [2] http://www.johnkay.com/2005/09/27/airlines-and-the-canine-fe...

Notice that avoiding that the plane disappears does not save any life. That's why people do not care so much about it, and that's why they stil disappear.

That said, good telemetry brings economical gains. Planes will stop disappearing in due time, just not by state decree, and probably because of more usfull devices than this one.

It doesn't save lives on the disappeared plane, but it may save lives in the future. If, for example, the cause of MH370's disappearance is something that could happen to others, and can be remedied at a reasonable cost, being able to find out what that cause is and how to remedy it would save lives in the future.

A low priority, but not nothing.

> Notice that avoiding that the plane disappears does not save any life.

Analyzing the causes of plane crashes saves lots of lives. You can't do that with a disappeared plane.

You usually get the black box upon recovery, which doesn't save the passengers, but does get you all the valuable data.

> You usually get the black box upon recovery, which doesn't save the passengers, but does get you all the valuable data.

And recovery is only possible if you can find the wreckage. MH370 is still missing, and as such the cause of its crash is unknown. If we had real-time GPS monitoring of airliners, chances are we'd have wreckage in-hand.

MH370 was rare. You typically have a small search space with a downed aircraft, and the black box announces itself acoustically underwater for up to 30 days.

http://en.wikipedia.org/wiki/List_of_unrecovered_flight_reco...

17 unrecovered black boxes since 1965. I don't think that calls for spending tens of millions of dollars considering ADS-B and Iridium will have the problem solved in the next few years: http://www.aireon.com/AboutAireon/AnIridiumInnovation

Missing a few blackboxes does not make all that difference in practice. Whatever happened with the plane was probably related in other flights already, and if it's really something that never repeated anywhere, it's not worth protecting against anyway.

s\related\reported\

"Hey, a plane went down right here. Want to send all your boats to see if anyone survived?"

Great! Get that boat there (in the middle of an ocean) in less than 16 minutes, or everybody will be dead anyway.

... ok. There are a LOT of boats in the world. How long could it possibly take to get the nearest one anywhere a plane goes down, anywhere in the world? Five minutes? Ten?

Well, sorry about the uninformative answer. I was in a rush...

Anyway, 6 minutes (not 16) is the time people are usually expected to still be alive in a plane after it gets on fire. Airports spend a lot to get cars on their area fast enough for rescue, there's simple no chance boats can get anywhere in the oceans that fast - when a plane falls on the water boats need days to get there. And all "the nearest boat" will do is catching fire together with the plane, you need well equipped boats with trained people.

Anyway, I'm having a hard time trying to imagine a scenario where airplane tracking fails, but people are still alive. I'm about to claim that's currently impossible, but that'd be just asking for a counterexample to happen and prove me wrong... When planes disappear, it's because they were either completely destroyed in air or hit the ocean at a huge speed.

"...a six-pound tracking system, about the size of a hotel safe, installed in the planes’ electronics bays...if something goes wrong — a sudden loss of altitude; an unexpected bank; engine vibrations — the system begins transmitting data to the ground, via satellite, every second. That six-pound box spits out reams of performance data, as well as the basics necessary for a search-and-rescue: coordinates, speed, and altitude....installation alone runs about $120,000 per plane..."

Should this cost $120k? I'm no expert, but that seems steep. I assume this data is already being written to the black box, so piggybacking off that data feed seems like it would be a pretty simple thing.

Maybe you're paying for the guaranteed access to the satellite network when it's needed rather than the tech on the plane itself. ....and of course a bit of a "no competition so deal with it" charge.

My neighbor growing up owned brokerage for refurbished aircraft parts (apparently there is a niche for everything). They'd refurbish some stuff in house but farmed a lot of it out. Anyway, one day he was giving me a tour of his office and they had a garage for storage in the back. He picked up this valve off the shelf, it was the size of an 80 count aspirin bottle. He said he bought it for $1200, spent another $900 to refurbish it and he was going to sell it for $5000. Absolutely mind boggling.

This is aviation engineering where everything needs to be checked and verified multiple times by multiple people because otherwise planes fall out of the sky.

Companies exist to make money, the market is very limited, the sales process is long etc. but feel free to compete.

People generally dont try to replace a 100k job with another one when they start companies, but are looking for a (multi) millionaire upside instead

Airlines and regulators have no interest in slapping on crude, consumer grade transmitters to existing aircraft. As is the case with any aircraft part, it needs to be a high-grade, reliable, tested, and certified.

When you start tying in transmitters to the aircraft's existing avionics, the need for a redundant, extremely reliable, tested and fault-free system is even greater. An absolute must.

That's where the big expense comes in. And I believe that's what passengers have come to expect. Personally, I'd rather fly on an airliner without a satellite tracking system than fly with one that hasn't been subjected to the same testing and certification standards as the rest of the plane's equipment.

The cost is the prohibitive factor. I can imagine that Delta, for instance, has no interest in a $90M charge to retrofit their aircraft with a system that they will probably never use.

The money would probably be better spent speeding up ADS-B transmitter installations, as well as incentive payments to Iridium to get their new satellites with ADS-B receivers into orbit. No point in reinventing the wheel.

I'm sure it's something that will be worked out in time. Perhaps the manufacturers themselves will be more interested in rolling out tracking technology in new aircraft versus airlines retrofitting existing planes.

But it's not an industry that responds quickly due to media hysteria surrounding rare and isolated incidents, that's for sure.

This is already being done, by Aireon.

Right, which is a subsidiary of Iridium:

http://www.aireon.com/AboutAireon/AnIridiumInnovation

And yet... and yet, the pilots I know who fly big jets brought their own GPS units into the cockpit decades before 'reliable tested' versions were certified. And used them. Because they were helpful, but not necessary.

And they are free to do so because the GPS units pose no risk to the aircraft, just as I use my iPad or smartphone on a plane.

Once you start talking about tracking systems that are tied into the plane's avionics in order to deliver information about the aircraft itself, that's another story.

I imagine partly because the airline was buying from the "11-20" quantity column. If Delta were to buy, with 700+ planes, they'd likely get a pretty good quantity discount.

A lot of the arguments in this particular thread are based on the assumption that an aviation grade (non-firmware locked to low speed / low altitude like some 90s models) an an antenna are somehow more expensive than $120K or more specifically that it would be more expensive than screwing around with the existing black box and flight avionics.

Making it all even weirder, lets price out something that does an even more complicated task, a TCAS unit that ties into the avionics and A/V headsets and sometimes HUD and the transponder and in real time warns of potential collisions. $15K to $20K installed. But this vastly simpler project is supposed to cost about ten times as much. Don't think so.

Also a lot of confusion about what a black box from 1980 vs 2015 actually records and what is necessary. Yes from a safety perspective a simple location tracker does nothing and you'd need a full black box feed. A classic programming example of not defining your goals means a fail. Have to spec if you just want a lat/lon or if you want a full engineering telemetry feed.

Finally, if the engine controller fails, and you're over an ocean, then you die. Pretty simple. On the other hand, 99.9999999% (seriously) of flight hours, if this flight recorder fails, much like a tcas blowing a fuse, you simply fix it before the next flight. A big old "whatever". Its not like most planes disappear on a regular basis or most planes are involved in near collisions. The MTBF of the all hardware is shorter than the mean time between accidents. So consumer grade is good enough as long as it doesn't burst into flame like a consumer laptop.

From the article: "In discussing about potential safety advances, Tony Tyler, the International Air Transport Association director general, said earlier this month that 'even though aircraft cannot be tracked in all cases, flying is safe. Over 100,000 flights operate safely every day.'" That's important baseline information. I would much rather fly on any of the airlines that serve the major airport in my town than drive here, and that's even though some website has just identified the drivers in my state, incredibly, as the safest drivers in the United States.[1] Airline flying as it is currently regulated and implemented daily is an astounding safe way to go to places near and far.

[1] http://bringmethenews.com/2015/01/04/study-finds-minnesota-h...

No matter how safe something is, we should still be trying to make it safer.

Rubbish. Perhaps you meant to say "Unless something is already completely safe, we should always be open to analyzing the benefits and costs of possible safety improvements, while being sure to spend the majority of our resources on areas that offer easier or larger gains".

> Any new technology also requires new training of pilots and ground crew, and it opens new avenues for mistakes.

I think this is the key point. This isn't iPhone level technology, this is closer to Space Shuttle technology, where things need to be tested to the extreme, radiation tolerant, and able to withstand the force of a literal plane crash.

Meanwhile they allow hundreds of relatively high power RF transmitters (passengers' phones) onboard. I call humbug.

Right, my point is that this new device has to be tolerant of radiation from passengers' phones, increased radiation from space, etc

And what happens if it isn't? And how would be this be any worse than no device at all?

The device recording plane data is presumably connected to important plane systems to record them. Can data/electricity only flow one way? Are you sure?

Why? for location data all they need is GPS. If you know location you find the black box which has the telemetry data.

Is there a reason a location tracker can't be self contained with it's own power supply and no connections to a plane's telemetry feed?

If you do that, the next time there's a crash like Air France 447, you'll see rending of garments and crappy articles like this (again) saying "Why have a black box when you can just transmit the data?!?!?!"

http://en.wikipedia.org/wiki/Air_France_Flight_447#Underwate...

> French Bureau d'Enquetes et d'Analyses (BEA) chief Paul-Louis Arslanian said that he was not optimistic about finding them since they might have been under as much as 3,000 m (9,800 ft) of water and the terrain under this portion of the ocean was very rugged. Although France had never recovered a flight recorder from such depths, there was precedent for such an operation.

Crashed: June 2009

Black boxes found: May 2011

Notice that the AF447 had great telemetry reporting its position right before the accident. Searching for things at the sea bottom is simply hard.

My phone doesn't connect to the aircraft's avionics. Big difference.

Not directly. But the field emanating from your phone will induce a current in the wiring of the plane if it isn't properly shielded to account for that. So even if there is no physical connection there could still be a resulting interruption in communications. All it takes is a crappy or un-intended ground point somewhere along the signal path or at either end.

One way in which I would stress test new electronics designs for resistance to such interference was to stick a mobile phone on top of it and then call that number. Worked like a charm most of the times: instant crash.

Which they didn't allow until recently. And of course a system like this would almost certainly be tied into the airplane's avionics and communications systems, and possibly also its power system. It really does have to be tested more than one might expect.

I don't think this expensive solution is the answer. All airliners already have ADS-B transmitters...the problem is that over the ocean there are no ground receivers to pick up the signals. Aireon is launching a set of satellites that will pick up ADS-B signals from all aircraft, and will offer the service completely free of charge to any airline in the event they lose contact with a plane.

I'm curious why a simple "Where am I right now" item that doesn't have to plug into the avionics and could possibly even have it's own power supply isn't sufficient to cover this. The biggest problem for both of those planes seems to be that we can't find them.

Figuring out why they crashed is possible once we find them and can get the black box. The only really missing component is the location ping and it should be possible to completely isolate from the plane shouldn't it?

You can buy a box that attaches to anything you'd like to track. For example with magnets to a steel shipping container. It has its own battery and gps and a satellite phone. It notices from the gps when the container is moving and starts sending position updates with the satellite phone. You can specify the interval, that then decides how many years the battery lasts.

So why isn't this a cost effective approach to the problem for airlines? You could even have a pool of them that you attach to airplanes that on overseas duty only to limit the costs.

The ultimate system would ping home every minute, not wait for a fatal even which may cause loss of power or antenna.

Certainly satellite data cost has come down to the point where you can track such expensive planes in real time?

Reply on news.ycombinator.com