Chomolungma
June 12, 2009
'Black box' obsolete
PARIS - WHETHER or not the black boxes from Air France flight 447 are found, the crash has shown that new technology is needed to record a flight's last moments in real-time, an aviation expert argues. Former Air Canada chief executive and ex-head of the International Air Transport Association, Pierre Jeanniot helped pioneer flight data recorders 40 years ago but now says the 'black boxes' are obsolete.
Now, maybe I am nuts and maybe I'm not. Bit CVR and FDR tools record continuously overwriting the old data with new data continuously. You get a couple hour record or more. The bit about wait until we're in trouble to transmit does not tell a pickled thing about how they got there. What would be needed is bandwidth sufficient to squirt up the data from the last couple of hours and then continue the data transmission until conditions return to normal or LOS (Loss Of Signal.) That initial squrt of a couple hours of data would require a TV signal sort of bandwidth to make it happen fast enough, and even then it might not all get there. Then maintaining the flow would require a narrow voice channel and a slow data channel at most. So you'd be wasting an expensive TV bandwidth channel. But it's wasted anyway because it has to be available at all times. And to make sure the most valuable data gets there first you'd have to use uncompressed data or data with a compression format that works in reverse. Send the newest data first working back to the oldest data. Ugh - the solution seems ugly to me as a communications engineer.
What MIGHT help for locating FDR and CVR is a means of automatically triggering a stream of flight data including the navigation system's estimates of position, velocity (speed and direction), and flight control settings. Cockpit voice can wait for recovery. The data would be designed to put the search as close as possible to the right place.
As many others want to do if we elected to send CVR and data with enough fidelity to get the voices well but compressed almost to death there is a good algorithm that gets it all through a 2400 bps (300 BYTES/sec) or so channel. The algorithm even tends to mute ambient noise. Somebody else mentioned the FDR stream is about 2 k BYTES a second or so. So you get a total bandwidth of about 2 to 3 k BYTES per second per plane with 10,000 planes in the air. That's an agregate data rate of 20 to 30 megabytes per second or a good two digital TV channels worth - if it's all collected and transmitted at one place.
Planes aren't all neatly packed in one place. So each is transmitting its own data. That means each one uses about 10 kHz of bandwidth, continuously, if you want reasonable error rates for a first cut round figure. That's 100 megabytes bandwidth, a piddly 17 TV signals wide or so.
Off hand that sounds terribly expensive. I'd stick to simply trying to make finding the recorders a little easier. And even that may not be worth it for a deep ocean recovery once every few years. The $100 million or even gigabuck SAR may be cheaper in the long run.
{o.o} (It would be fun to have that much bandwidth to play with. But it might also be fun to have that much money to play with.)