Modular Airports

[Wezz6400]

As you can see here Schiphol does in fact have 6 runways. The 6th one is quite new, which is why the sattelite pictures (zoom out a bit further) do not show it yet. Oh and don't get confused with the two motor ways and the railroad going through there as well.

[Rubidium]

All six runways of Schiphol are shown at that satelite image.

[AlexW]

I would love to see Heathrow, having spent far too much time there...

[richk67]

I would love to see Heathrow, having spent far too much time there...

It looked better in 1955 than it does now (basic 2 runways) ... or even in the future plans.

[Wezz6400]

All six runways of Schiphol are shown at that satelite image.

In fact the link I posted is made from an airplane, if you zoom out one level you'll get the sattelite picture and the runway on the far left will "disappear".

[SuperTycoon]

A more general question:
How are these new airport designs being made and coded? The prospect of designing some airports seems attractive, but how much knowledge is required for coding?

[richk67]

Take a look at the NewGRF_ports branch on http://git.openttd.org. This shows the latest commits.

I can currently place an airport with state machine and graphics via newgrf. I am just sorting out rotation, and then need to sort out save and load.

Although it is an old design (commuter), it is enough to prove that it works, and more complex ones can be added later.

Is it easy? Not really. Designing a non-deadlocking FSM takes quite a bit of thought, and is definitely a paper-first exercise before trying to code it into newgrf (or directly into OTTD for that matter).

[MeusH]

Take a look at the NewGRF_ports branch on http://git.openttd.org. This shows the latest commits.

Excuse me changing the topic, but is GIT similiar to TRAC?

[athanasios]

Munich is a monster. 12 terminals, 3 helipads, 3 runways... and at full loading it can have 18 aircraft in motion on the ground at the same time.

But it is just a toy compared to real life ones. (Nevertheless I would like to have your concept of Munich in OpenTTD. )
I can count 24 terminals in our city's airport, only for passengers. Light aircraft and cargo aircraft have separate. And I think there are 6 helipads. See posted images. Map is not accurate regarding terminals. (Just a map, missing many details. Even rail/metro station is missing. I posted this map to see how many things are missing from game airport.)

Only 1 aircraft in taxiway in modular airport? You' d better look at this jammed taxiway (Pity, if I knew I 'd go and take some shots):
http://www.planepictures.net/netshow.php?id=628470

Yesterday we had >600 flights. Luckily my auntie travelled last Saturday. Even then airport was packed. Propably yesterday it was hell.
Landing phase in Airports has to be changed. An airplane will start landing and the ones following turn around. It is not real.

[richk67]

OTTD is a game, not a simulation. It doesnt matter whether it is "real" or not... its whether it is programmatically safe. Which it is.

Anyway, IMO it is realistic to an extent. I have been on an aircraft that got within 1 mile of the runway, before it aborted and went for a go around because the pilot didnt like the separation he had to the preceding aircraft. At some point the pilot/OTTD has to decide "land, or go around". That is what our state machines do.

[Ben_K]

I can count 24 terminals in our city's airport, only for passengers.

By that, do you mean 24 air bridges? I dont think there are many airports in the world with 24 terminals. It would be a passenger nightmare! In which case, at work we have 63 stands served by air bridges, many of which can be used by 2 aircraft at once with individual air bridges!
Photos: Picture1, Picture2

...Only 1 aircraft in taxiway in modular airport? You' d better look at this jammed taxiway...

Im guessing you know that they are parked. None the less, its impressive use of space

Yesterday we had >600 flights.

Thats an impressive number... although, we had 2 days last week with 874 movements. The record for the busiest single runway airport in the world (which we hold) is 877 movements. And that is with 1 single runway!
Sadly, as much as Id love a Gatwick copy on OTTD, the system would never make it a worthwhile airport to use.
(Unless rich can programme reduced seperation and conditional line-up clearances )

...At some point the pilot/OTTD has to decide "land, or go around". That is what our state machines do.

I must admit, it would be nice to see the aircraft make a lower approach before going around - maybe in the tile before it lands? Then maybe turn away during the missed approach. Would look more in-keeping with High capacity airports.

[MagicBuzz]

I'm just thinking about something Richk67...

Do you know Microsoft Flight Simulator ? And... AFCAD ?

This tool allows the user to design the ground pattern of an airport, with only a few clicks :
- Runways (you can even say "this one is for landing and this one is for takeoff")
- Taxis
- Holding positions
- Parkings

Then you can just load them in Flight Simulator and even send the AI aircrafts on it, they will know how to use it : a single aircraft per runway, do not collide on the taxies crossways, and don't try to land if there isn't any free parking (in fact, it will do like in reallity : it will start the approach, then abort it if no parking is free).

Regarding the FS AI, there is also a nice idea OTTD should clone : right now, the first plane that enter an airport area will make all the other aircrafts to wait, without taking attention of the logic of an airport. By this way, we can see 10 planes flying arround an internationnal airport while only 1 parking is occupied, because a plane is heading to the runway : flying aircrafts are always with the highest priority, a plane on the ground can just stop and wait while the flying ones can't.

So, what about :
1/ Ginving all planes that are doing stacks arround the airport a higher priority : not any plane on the ground can enter the runway till there are planes on approach and still free parkings on the ground. The best thing should be making two steps during approach : planes doing stacks will start the first approach step even if there is a plane on the runway -but not if not any taxi is free- then enter the 2nd step only if the runway is free. By this way, approch is really faster
2/ Maybe creating a (very) simplifier AFCAD and allowing OTTD to use it (ok, I know it's not easy at all) then people could design their own airports easily.

AFCAD file sample (Luton - EGGW)

You can see several things :
- Black lines : runway path. You can see there are black loops at the end on the runway. It's because a plane that is at the end of the runway and lining up must be considered as on the runway
- Green lines : taxi ways.
- Red dots on the taxis : hold on. When a plane comes from the parkings, it will stop there till there is a plane on the runway of approaching. If it comes from the runway, the runway is marked "free" when the plane passed the red dot.
- Yellow dots on the taxis : safety hold points : only one aircraft can move at the same time between 2 yellow dots
- Some parkings (actually they are gates) are overriding. FS can detects this, and disallow the use of the overrided parkings when a big plane uses one, and forbid a big plane to use a small parking. This should open the airport layout for the 32 bits graphics that uses variable size for vehicles : two Douglas MD-11 can stand where one Airbus A380 can't.

I think it should be a good idea to let the player designing these paths and signs like railways and using different colors as well.

[Born Acorn]

AFCAD file sample (Luton - EGGW) :
LARGE IMAGE

Can you not break the tables and turn that image into a link or resize it please. Table Breaking is not tolerated.

[athanasios]

Ben_K:
* I used the word termimals in a similar way richk used it when describing his airport to avoid confusing him. My mistake. I would confuse millions of passengers. Thanks for correction.
So we have 1 main passenger terminal and 1 satellite terminal with 24 bridges, and many many more stands without bridges for light aircraft and cargo aircraft ('cargo' terminal?).
* About parked airplanes: May I suppose they are waiting their turn for an empty stand? Even if not, it is something to be considered instead of parking 10 planes inside a single or 2 depots. A suggestion that is realistic. (Others have suggested something similar for trains a long time ago.)
* 877 movements with only 1 runway? I simply can't believe that! You beat us! We got 2 records too! European Airport of Year 2004 (British Transport Management Institute) and the other one is a negative one: most expensive airport in Europe.
Ben_K / richk:
* I like such big airports to be in game. Even your favorite airport (Gatwick) Ben_K can be efficient if airplanes don't make that nasty turn when another plane is landing. In real life you can see airplanes lowering to land many many kilometers away from the airport, one behind the other, and consider that the runways are 4 km in length. When something goes wrong they can even touch down and take off: This I suggest should be added to OTTD. With current airports of 6 tile runways it is useless but with bigger onces like yours richk it makes sense as it will increase efficiency dramatically.
Maybe I am dreaming too much. Planes and ships still are not conscious of their position and pass through each other > They travel in 4 dimensions!

[Ben_K]

* About parked airplanes: May I suppose they are waiting their turn for an empty stand?

Nah, they are parked there. You can see because of the orange cones to make sure people stay clear. Also, they are parked right up to the edge of the taxiway very close together. They would probably be towed from there to an empty stand wwhen its available.

* 877 movements with only 1 runway? I simply can't believe that!

Honest guv'nor! Most of us here suspect we will break that this year and there are rumours we may have a busy enough day to get close to 900.... we'll see!

I like some of your ideas for the airports although sadly I expect most of it is wishful thinking. It just depends how in depth Rich can make his system (Which Im sure, if he can, he will!)

[richk67]

Sadly, as much as Id love a Gatwick copy on OTTD, the system would never make it a worthwhile airport to use.
(Unless rich can programme reduced seperation and conditional line-up clearances )

Not quite sure what you mean by conditional line-up clearances, but the system does allow an aircraft to release blocks explicitly, so on my Munich, I can have a long line of aircraft queuing for departure. SFO is even more fun, with dual parallel landings and takeoffs.

...At some point the pilot/OTTD has to decide "land, or go around". That is what our state machines do.

I must admit, it would be nice to see the aircraft make a lower approach before going around - maybe in the tile before it lands? Then maybe turn away during the missed approach. Would look more in-keeping with High capacity airports.

Ive not fully completed my changes to the height-change system, but you now specify in the state machine movement what height each position is at. Thus, in theory (Ive not written one yet, but it should be possible), you could program a spiral holding stack.

Normal flight level is 80, but you could set it for say 60 on the "Flying" state - where the decision to land is made. Aircraft going on to land will pick up the new height of the next position (0 if next position is end of runway). If landing decision fails, the height of the next position (back in the circuit), will be 80 again, so it will climb away from the aborted landing.

Currently up/down movement is strictly 1:1. I want to change this to being dx/dy, so allowing for whatever rate of climb you want. Kaitak with its near vertical approach or O'Hare with its looooooong approaches (with lots of a/c in the approach).

This is the beauty of the state machines - you can split a flight into as many sector blocks as you require, and strictly control how other aircraft react. Usually, our current holding patterns have no blocks as this would limit the number of aircraft that can circle. But in theory, you could have a 10 plane approach line, and if all the landing approaches are full, then the excess is bumped into a pre-landing holding stack.

Your imagination is the only limitation... (OK and a ton of state machine skill )

[richk67]

Ben_K:
* I used the word termimals in a similar way richk used it when describing his airport to avoid confusing him. My mistake. I would confuse millions of passengers. Thanks for correction.
So we have 1 main passenger terminal and 1 satellite terminal with 24 bridges, and many many more stands without bridges for light aircraft and cargo aircraft ('cargo' terminal?).

I just re-used the terminology used in the code. I might rename them however, as the FSMport scheme can apply to ships and road vehicles too. (In theory, but its a pretty solid theory). I was thinking of either berths or bays. I agree the terminal is the building. If anything, terminal may be an appropriate name for what we currently call a terminal group.

Stands: too a/c specific.
Berths: OK for a/c & ships. Not sure about road vehicles.
Bays: Not so good for a/c. Good for ships. Very good for road vehicles.

When something goes wrong they can even touch down and take off: This I suggest should be added to OTTD. With current airports of 6 tile runways it is useless but with bigger onces like yours richk it makes sense as it will increase efficiency dramatically.

For most of that, all that is required is segmenting the runway into multiple blocks, and releasing them in sequence. You can already program it that if the next block is blocked, rather than pause, the a/c can choose another matching command further down the sequence. So you could have:
44 42 41 21 //if landing (44), block 42 is clear, release block 41 (behind you), reserve block 42, and proceed to position 21
else
44 00 41 25 //if landing (44), (to get here, first test must have failed), no block test so always OK to move, release block 41, proceed to position 25.

The *only* thing the current system is missing is a "clear current heading" instruction, since with the previous example, the aircraft will continue to have the 44 landing heading. I think I may introduce a new instructional command: FF 00 00 nn meaning FF-clear current heading, and change it to nn. So in the example, position 25 would read:
FF 00 00 7F //clear current heading, and set heading to FLYING.
00 00 00 26 //for all headings, move to position 26

Powerful stuff

[Ben_K]

Not quite sure what you mean by conditional line-up clearances, but the system does allow an aircraft to release blocks explicitly...

A conditional line-up instruction is "After the landing a/c on finals, line-up RWY XX". What then happens is one plane is lined up ready to roll as soon as the previous one vacates. Rather than waiting until one is clear before the next one goes. It is one of the key elements of HIRO (High Intensity Runway Operations), especially on a single runway like ours (Gatwick).

...lower approach before going around... ...a spiral holding stack...

By lower approach, I meant over the tile before the runway, so that its a real last minute yes/no decision. Its just what Im used to here (Especially today!! )
As for a racing track hold, that would be really awesome to see! And presumably we would be able to have multiple holds per airport, like Gatwick or Heathrow?

This is really intriguing work you're doing Rich. Im really quite impressed with the sound of it. Anything I can help with, just let me know!

[richk67]

Well, the power has always been there; its just been hardcoded, and so not really accessible by those afraid of C++

There are only really two major changes I have added; each aircraft owns its blocks - previously only the airport owned blocks. Secondly, there was no explicit and reliable way of releasing blocks, so the state machine could end up in a real tangle when you tried to add more complex actions.

Other than that, most of the changes are just about making it load via newgrf; thus providing easy custom graphics.

As for speeding up takeoffs - you could easily have it that once the taking off aircraft has got 2 or 3 tiles down the runway, it releases the start position tile. The new aircraft would proceed there and wait for the other a/c to clear whatever is required... say on tile 4 of the runway. Then it could start its takeoff.

On landing, yes, you could have a position just before the landing tile, say at altitude 16, and if the runway runout is not clear at that point, then abort and climb away. Easily done in the new system. I'll add the "clear heading" function tonight to really make this possible.

[MagicBuzz]

Richk67 & Ben_K > You land several planes at the same time on the same runway in your countries ?

I can understand a few planes can approach at the same time, but they must engage fail approach procedure on final if a plane is still on the runway, even if it's at the opposite side...
What could happen if the already landing plane drop pieces of metal on the runway ? What could happen if the second plane have a brake issue while the first one is stuck at the end of the runway ? When you see a plane can just burn several minutes after a RTO, taking these risks is just suicide... (remember the A340-600 RTO test reject video).

The only exception I know in France is on small airports that uses only one runway for landing and take off : a plane on the ground can enter the runway as soon as the landing plane passed its position.

PS : LFPG has also one record. The shorter airport installation service before causalities due to building decay.


Born Acorn > Sorry, I didn't noticed the picture was too big on my wide screen