Baggage Handling System – Schiphol Airport [video](youtube.com)
youtube.com
Baggage Handling System – Schiphol Airport [video]
https://www.youtube.com/watch?v=rv-Y0-ruzi8&feature=youtu.be
7 comments
Is there a version without the edits? Because I found this practically unwatchable; it kept cutting away just as things started to get interesting.
This video of the T5 system at Heathrow is better and shows equipment just as cool (though no robot at the end)
https://www.youtube.com/watch?v=Wn8qogHH9bM
Oh and if you want a REALLY crazy version of the Schipol video, this one is actually a 360 degree live draggable version of it - no kidding!
http://www.schiphol.nl/Reizigers/OpSchiphol/Bagage/BagageVid...
https://www.youtube.com/watch?v=Wn8qogHH9bM
Oh and if you want a REALLY crazy version of the Schipol video, this one is actually a 360 degree live draggable version of it - no kidding!
http://www.schiphol.nl/Reizigers/OpSchiphol/Bagage/BagageVid...
The first one isn't an FPV... and the second one is also full of edits!
Was looking for some backgrojnd and found an overview presentation [pdf]: http://netlipse.eu/media/77918/11nwm-bratislava-lex-pepping-...
Wow. It's actually a miracle that this works most of the time. That suitcase is pretty standard, I could imagine a million ways how some backpack or other unusually formed piece of baggage would jam that thing.
And how on earth do they route individual packages through the system with that speed? Is there some pattern recognition system that identifies them on the fly and then routes left, right, left right, etc?
This reminds me of a funny part of a book I read about automation (and its impact on human labour), which described how modern corn mills identify bad grain: they shoot every single grain through a high speed tunnel and identify malformed (bad, moldy) grain with pattern recognition and then filter them out.
And how on earth do they route individual packages through the system with that speed? Is there some pattern recognition system that identifies them on the fly and then routes left, right, left right, etc?
This reminds me of a funny part of a book I read about automation (and its impact on human labour), which described how modern corn mills identify bad grain: they shoot every single grain through a high speed tunnel and identify malformed (bad, moldy) grain with pattern recognition and then filter them out.
They have sets of barcode scanners dotted around routing points. With enough of them and the big barcodes on the baggage tags, they can scan most automatically. Any that fail scanning are diverted to a side belt where a human scans them with a handheld unit.
That was incredible. Remove the "flip aside" parts and it's much better than your average rollercoaster! :D
This is unbelievable. How can they possibly justify the cost of this system?
Same as for any automated system: reliability and throughput.
Some of these airports are enormous, Schipol especially so. They're handling 150,000 bags per day at peak times. If you generously assume a 1% error rate (http://panko.shidler.hawaii.edu/HumanErr/Basic.htm) and that each bag is touched by only one human, then you'd be misrouting or losing 1500 bags per day, which is huge. But of course bags aren't going to be handled by just one human, there would be many making complex routing decisions, over and over again. So the true loss rate would be much worse.
Bear in mind that if the airline doesn't get the bag to the right place then the bag has to be sent onwards later, and then the airline has to pay for home delivery of the baggage. So the cost of mistakes adds up fast. Then you have the sheer amount of manpower needed to move all the bags around by hand. And not just any manpower, it often has to be literally manpower because you need strong men to do this as the bags are so heavy, and they can't work long shifts because they get tired.
It's pretty easy to see how such things can be justified. You wouldn't be able to scale air travel to the current levels it sees without such systems.
Some of these airports are enormous, Schipol especially so. They're handling 150,000 bags per day at peak times. If you generously assume a 1% error rate (http://panko.shidler.hawaii.edu/HumanErr/Basic.htm) and that each bag is touched by only one human, then you'd be misrouting or losing 1500 bags per day, which is huge. But of course bags aren't going to be handled by just one human, there would be many making complex routing decisions, over and over again. So the true loss rate would be much worse.
Bear in mind that if the airline doesn't get the bag to the right place then the bag has to be sent onwards later, and then the airline has to pay for home delivery of the baggage. So the cost of mistakes adds up fast. Then you have the sheer amount of manpower needed to move all the bags around by hand. And not just any manpower, it often has to be literally manpower because you need strong men to do this as the bags are so heavy, and they can't work long shifts because they get tired.
It's pretty easy to see how such things can be justified. You wouldn't be able to scale air travel to the current levels it sees without such systems.
not sure if this is a new system but I flew through Schiphol a lot around 2010-2012 and the baggage was very slow. Typically half an hour wait. Whenever possible I flew with carry on as I found it frustrating to have an hour flight then half an hour wait.
My employer at that time was using MS Excel to parse the log files of all sensors of the predecessor/older part of the system that is shown in this video. I told her that I could probably do it 10,000 times faster by creating a simple program in Visual Basic.
Visual Basic was way too slow on our "high-end" Pentium laptop to parse that many 1MB log files, so I rewrote it in C++ (learned it on the job)[1]. The managers got insight in the performance of the various components of the system which they never dreamed of having (they were hardly aware that the system was creating this detailed log messages).
By the way, the system is developed (at least partly) by Vanderlande [2].
[1] Took me a couple of days, probably even weeks back then. I would write the code in 1 or 2 hours now.
[2] https://www.vanderlande.com/