Hacker News new | past | comments | ask | show | jobs | submit login
What happens when an air traffic controller is asleep? (law.harvard.edu)
127 points by yan on April 14, 2011 | hide | past | favorite | 22 comments



http://www.avweb.com/news/sayagain/191072-1.html has a (now retired) US air traffic controller's thoughts on a fatal collision involving two of the three planes a European controller was responsible for. It suggests that not all cases of sleep deprived controllers are harmless.


Like most serious accidents, the sleep deprived controller is just one part of what went ridiculously wrong there. Had the controller actually been sleeping the entire time, and the pilots followed the automated systems, it's entirely possible that they would have been ok. In addition, the equipment failure/downtimes severely limited any operator's situational awareness.

It's one of the cruel ironies here, that a completely passed out controller likely would have lead to a better outcome.


In a macabre follow up to the crash, the (Danish) ATC on duty was murdered by a Russian man who had lost his wife and children in the crash. 3 years later the man returned free to Russia, was hailed as a hero and given the position of a deputy minister in North Ossetia.

http://en.wikipedia.org/wiki/2002_%C3%9Cberlingen_mid-air_co...


Exceptional link, could stand on its own. Thanks!


In olden times (yes, this is anectodal), the process of producing nitroglycerin would be supervised by a person sitting on a one-legged chair, so that he would not fall asleep.

Source: http://bit.ly/hO5N1s (translated from Swedish)


This article is spot on. An excellent description of the way traffic flows through airspace and the role of ATC in movement of aircraft. I can add that as a pilot is would be a bit disconcerting and unusual to be on an instrument approach into an airport and be handed off to the tower only to receive no response. This would likely happen with only a couple of minutes of the flight left and at a time when workload is high (assuming IMC conditions). There would probably be a little confusion before deciding to switch back to the previous frequency.

Not particularly dangerous - but the pilot does have to be prepared for handling some atypical tasks on an otherwise routine approach

Edit to note: VFR (visual flight - night or day) would be a non-event (from the pilot's perspective) with a sleeping controller so long as traffic is relatively light, as it would be at night most places


Yes, for a nearly deserted airport the tower controller isn't as important as people would expect.

The article doesn't mention it but I believe that the sleeping controller is also supposed to be handling ground control - planes taxiing and ground vehicles like fire, security, baggage, etc.


It's short and it's definitely worth a read. I had no idea that there were airports without control towers.


I worked for 18 months as a fire & rescue person at an airport. While it was a real airport (terminal building, check in desks etc) it was a small regional airport, and thus we typically had only 1 or 2 scheduled movements per day.

We had one ATC at the airport (they came and went in 3 month rotations). They were in the tower most of the time monday to friday (during normal office hours) and obviously there for all scheduled flights, day or night - but equally he didn't feel the need to hang around. Plus of course toilet excursions.

We had a traffic-control radio in our office, so we could hear the movements and (on weekends for example) issue a simple phrase like "good morning echo-foxtrot, this is xxxxxx, proceed with unmanned tower approach".

Our section was obviously manned 24/7 so in theory we could hear distress calls as well, but I don't think we ever did. One helicopter made an approach over the hangar, and landed on the hard-stand (not exactly what they're supposed to do) but apparently he'd pretty much run out of fuel. (According to the pilot he didn't turn the engine off after landing - he ran out of gas.)


Having a control tower doesn't mean much - on the Isle of Barra the airport does have a control tower but the runway is a beach:

http://www.youtube.com/watch?v=STqmbc8k9rU


There are also airplanes without radios (though such planes normally operate at untowered airports).


You'd be amazed where you're allowed to land a plane. Private grass field? Sure, but look out for rocks.


Indeed - Fraser Island off Australia has a (75 mile long) beach that is officially a public road, with the added regulation that starting and landing aircraft (which offer sightseeing trips) have right-of-way.


It might seem like controllers are not that necessary but (similarly to pilots of modern airliners) they are very important when stuff doesn't go as planned. Pilots can take over a and pull a Sullenberg, controllers can provide pilots with information on speed and height in case their instruments don't work (ground controlled approach style)



Surprised no one has mentioned this. Disclamer: ex-employer http://www.quintiq.com/news-and-events/news/2010/quintiq-lan...


One important note that isn't mentioned here. From what I understand, major/towered airports often have crews working the ground at night (due to the low traffic), towing planes, etc, and communicating with the tower over their own frequencies. Not expecting an unmanned tower themselves, they could easily try to "wing it" as best as they could to get their job done.

Sure pilots can avoid each other on the normal control frequencies, but the situation on the ground could be unexpectedly dangerous to ground crew/vehicles, especially in a night landing.


Hanscom Field is a bit smaller than DCA. A Cirrus is a bit smaller than a 737 (or whatever the planes landing at DCA were.) DCA is, oddly enough, immediately adjacent to Washington D.C. and Hanscom is hundreds of miles from D.C.

If the sleeping controllers were at Hanscom rather than DCA, this situation wouldn't have blown up nearly as far.


A Cirrus is quite a bit smaller than a 737; their biggest seats seven, and the smallest 737 built today seats over a hundred.


This article suits one HN guideline perfectly: "anything that gratifies one's intellectual curiosity".


It's really the difference between peer to peer and token ring.

One is done on a general channel which approaches failure the more traffic.

The other uses a "token" to signify communication and passes the token to whom they wish to speak to.


For some reason I feel pretty sleepy after reading this.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: