By Frederic Friedel
Planes disappear. In 1937 aviation pioneer Amelia Earhart went missing over the central Pacific Ocean while attempting to circumnavigate the globe. No remnants of her plane were ever found. In the late 1940s aircraft started vanishing in the “Bermuda Triangle”, leading to industrial level conspiracy theories. Planes going down over water are particularly difficult to find and the search for them monstrously time-consuming and expensive.
On 8 March 2014 Malaysia Airlines flight MH370 vanished on a scheduled commercial flight to Beijing, after performing a number of mysterious route manoeuvres. The search for the missing plane has been the most expensive in aviation history — more than a hundred millions dollars — and over two years later all we have found are a few pieces of matching debris, washed ashore on the African coast.
And on May 19, 2016, EgyptAir flight MS804 went down over the Mediterranean. After the first ten days some scraps of debris and some body parts were found, but there is still no certainty on what caused the crash. And it is vitally important to know — was it terrorism, was it caused by an explosive smuggled onto the plane in one of its numerous stops on the day of the flight?
Experts suggest that it will take months or even years to retrieve enough wreckage to determine exactly what happened. When Air France Flight 447 disappeared off the Brazilian coast in June 2009 it took search teams nearly two years to retrieve the black boxes from the ocean floor and another year to finally establish the cause of the crash — faulty airspeed indicators — and to hear the pilot’s final words: “Damn it, we’re going to crash”.
The key to finding a plane, and finding out what caused it to crash, is hidden in a flight data recorder, which preserves as many technical parameter as possible from the recent history of the flight, and in a cockpit voice recorder that captures audio sounds from the cockpit, including the conversation of the pilots. If these two hardened, water-proof devices are found intact, they can provide valuable information to investigators on what transpired in the minutes and hours before the crash.
Black boxes, which are actually bright orange in colour, are stored deep inside the planes and go down to the bottom of the sea with the wreckage. They are equipped with underwater locator beacons which are automatically activated and will signal for around 30 (currently being upgraded to 90) days under water. But it is of course notoriously difficult to retrieve black boxes when they are under miles of ocean, especially if the sea bed consists of rugged terrain.
So why are we still relying on a device and a strategy developed in the 1950s to monitor modern aircraft traffic? Isn’t it time to upgrade? This has been discussed — most intensely after a plane goes missing, and the general demand is that the data should be streamed live, via satellite, to recording systems on the ground. To which the conclusion inevitably is: too expensive, would cost billions to install and maintain.
There is a certain justification in this objection: there are over 100,000 commercial flights per day, which would result in hundreds of millions of hours of transmission and recording per year. Accidents almost never happen — if you took a flight every single day of your life you could expect to die in a crash in around 30,000 years. Streaming the flight and voice data of all commercial flights would be an unimaginable waste. And completely unnecessary. What you need to do is to have the critical (and extremely rare) information at your disposal immediately after an accident occurs.
Is this possible, and is it possible at a reasonable cost? Like $75 per plane? Well, as a thought experiment we could start with a hardware budget in this range. The innards of a standard low-cost Android phone will provide most of what we need: a wireless connection to the built-in black boxes of a plane, a micro SD card, and some attachment material. And a bit of programming — a few hours of it. The device, which I have called the “F-box” (for Flight Box) in discussions, receives whatever the black boxes are recording and stores the data in a loop — overwriting the oldest material as the memory chip fills.
The F-box is attached to the outside of the plane, ready to detach itself if submersed in water. It would be surrounded by flotation material (adding some more cents to its cost) that insures that if the plane hits the sea, and it was not directly destroyed in the crash, it will soon be floating around at the surface. If we want to stay within the $75 budget we will encase the box in Styrofoam and attach it to a nook in the plane using water-soluble glue — okay, that’s pushing it a bit.
Once the F-box is out there floating on the surface it would send a wireless ping every minute — or every ten minutes or every hour, to optimize battery life — with exact geo-coordinates of its location (did I mention that an Android cell-phone has built-in electronics to do exactly this?). Finding the exact location of the crash would be quite easy, and we would have instant access to the final hours of the destroyed plane.
But that is not the only safeguard we want to install. We will program the F-box to know the exact planned route, directions, cruising altitude and destination of the flight. This is something I can almost do today with a free app which gives me flight data when I point the camera of my smart phone at a plane flying overhead. So no great problem there. Using the built-in GPS the F-box tracks the course of the plane and makes sure it is not deviating from the flight plan.
If it is, immediate action is taken. Say the plane starts to descend, or veers off-course, or depressurizes — or say a panic button is hit by the flight personnel — then the F-box immediately starts to broadcast whatever it is currently receiving from the black boxes. It does this using whatever technology is available on the flight. This could be broadband WiFi, which many planes have now installed; radio; or best of all Iridium communication, which is globally available. Make a special deal with Iridium, which costs $49 per month for private users. This might add a bit to the $75 hardware costs, but would be well worth it.
Anyway, the F-box begins broadcasting whatever it is receiving from the black boxes, but it also broadcasts what it has recorded before the emergency action was triggered. And it does this backwards: the data is sent in reverse, and continues being sent until the box detaches from the plane and has only its battery left to draw power from. After that it is retrieval time.
The above means that when a plane goes down, like MH370 or MS804, we will have instant access to the final minutes (or hours, in the case of the Malaysian plan) of the flight, and to the exact location of where it all happened. We would not spend weeks (or months, or years) trying to find out where the wreckage is located, speculating if it was technological malfunction or terrorism, and sending tens of millions of dollars doing this.
I will admit the above plan is not fool-proof: the F-box could detach itself prematurely, and be lost (another $75 to replace it); or it could be completely destroyed at impact; or communications with the black boxes or satellites could be disrupted. All of this can be addressed and solutions found. The point is that we are talking about hardware costs that are equivalent to a short-haul economy class flight ticket, for the simplest possible solution. Give me a budget that is in the range of a first-class ticket and I will provide something completely foolproof — like building a little backward facing recess for the F-box that is not exposed to rain, or sensors that will detach the box if it is immersed in water. With other words: for a reasonable budget you get a solution that will fail only under the very direst of circumstances.
A number of brainy friends have sent me comments on the above thought experiment. John from Britain wrote:
I think the power supply to your F-box might be a problem. You commented on the WiFi (through the metal skin of the plane??) but didn’t discuss how the battery is kept charged. I don’t think that many plane designers would be happy to have a hole drilled in the side of the fuselage for a power cable to go through.” John thinks the simple fact is that my device would have to:
1) Stay attached to the plane in a 600 mph-airstream, through thunderstorms, turbulence, etc.;
2) Reliably detatch on entry into water;
3) Have a connection to supply power while it is attached;
4) Detatch this connection on entry into water;
5) Have a sufficiently low density to float to the surface.
It’s really hard to imagine how all this could be done and nothing which could accomplish it has been proposed.
Ken in California wrote (he abhors capital letters):
almost all commercial planes broadcast their position, altitude, temperature and wind every couple minutes. these are used for multiple purposes: 1) modern version of predicting winds aloft. old version (still used, but ignored in the u.s.) is weather balloons. 2) commercial flight trackers that have web aps that allow one to see where a flight is.
putting a much higher rate transmission of many more parameters would swamp any data channel. i think that allowing the box to self-diagnose a problem is not workable. only continuous transmission and/or recording will work. a possibility would be to transmit locally, and have near flights record. that means that the black box in a plane will record not only its own data, but data of nearby planes.”
Tommy from Hamburg said he had done a little googling and came to the following conclusions:
It appears that there are only about 20,000 planes in service. A large majority of those fly mostly over places like the US, Europe and East Asia, where they are never or only rarely out of communication range, so there isn’t a problem for those planes. Looking at flightradar24.com there are about 10,000 planes in flight right now, and I would guess that over 90% are over land, and most of those within visible range of a city. So that leaves probably well under 2000 planes out of direct communication range at any time. I was surprised how low all these numbers are.
According to this thread on stackexchange, planes already send out their location twice a second, but there’s not necessarily anything available to listen, to pick up the signals. Satellites that are launching soon might have receivers for the plane signals. But it appears that most planes don’t have satellite communication yet.
Considering the enormous cost of the MH 370 search (over $130 million), it might actually be worth retrofitting some satellite communication system to all existing long haul planes, especially since that would enable services that airlines want to be able to provide on planes anyway. The company Gogo seems to offer just that. Here’s their ground communication installation, and they also offer satellite communication upgrades. I suspect we will see these being installed on most planes soon anyway.
And I’m not convinced that the amount of bandwidth needed is all that much. In fact I’d go as far as to say that even if every single plane in the air transmitted many parameters every second and included a continuous compressed audio stream from the cockpit, the overall bandwidth would be less than 50 mbit.
Perhaps we should leave it all to Gogo, who need to add a recording device such as the one described above to the WiFi systems with which they are currently retro-fitting commercial airlines. This is more than the $75 I proposed (somewhat tongue-in-cheek), but it will also provide general connectivity on long haul flights. So it will be paying for itself.
Addendum: On January 17, 2017, after nearly three-year, the search for Malaysia Airlines Flight 370 was officially abandoned—forever. It had covered 120,000-square kilometer of ocean and cost $160 million, but has not been able to locate the aircraft.