Jump to content

Jim Vile's approaches for AIs


hjwalter

Recommended Posts

Hi Guys,

 

The late Jim Vile, may his soul rest in peace, has made some excellent approaches to FS9 airfields/airports so that AI aircraft do not visibly fly through mountains in their immediate areas, e.g. Innsbruck (LOWI), where such AIs approach very realisticly through a winding valley.

 

I've been busting my brains in trying to find out what the basics of Jim's methods were but even after disassembling some of his BGL files, I still have no idea how to even begin.

 

My interest in the above is for the small picturesque addon Ambrolauri airfield (UGAM) in a very mountainous area in the state of Georgia and which was not included in the default FS9. In the meantime I've edited the included "normal" approach BGLs for flyable (light) aircraft and ATC now vectors me in such a way that further visible and extremely spectacular approaches through the winding valleys, have become possible. However, AI aircraft sometimes beat me to it because they have "straight in" approaches through the mountains.

 

Advice anyone ?

 

Thanks in advance.

 

Hans

Link to comment
Share on other sites

Hi,

 

I have only used ADE to create such approaches, and will describe this using ADE.

 

First, AI aircraft will only follow a simple approach to the runway - complex approaches with transitions will not be followed.

 

They will use an ILS approach to the runway, but only the part of the approach on the main ADE page of an ILS approach - no legs, no missed approach, and no transitions.

 

The AI will:

 

1. Head to the waypoint displayed on the main page of the ADE ILS approach. It's listed as Fix Ident. It doesn't matter what direction it is coming from, it will head directly there.

2. Descend to the altitude in the Appr Alt Feet line, crossing the above waypoint at that altitude.

3. Depart that waypoint at the heading in the Heading Deg. line, descending constantly from now until it arrives at the runway.

4. When the plane crosses the extended centerline of the runway, it will turn to the runway heading and land.

 

So step 4 is how you get the "curved approach". In step 1 and 2 you set the location of the waypoint so you are at the starting point of your approach and at a reasonable alititude. You then set the heading in step 3 so the plane heads down the valley. Then in step 4 the plane turns from that heading to the runway heading and lands.

 

You cannot create more complicated approaches with more than two turns (turn 1 is the turn at the waypoint, turn 2 is the turn to the runway heading).

 

Hope this helps,

Tom Gibson

 

CalClassic Propliner Page: http://www.calclassic.com

Link to comment
Share on other sites

When I set an AI flight plan in hills and mountains I make it an IFR, even if the AC is flying low ATC will tell it to flyhigher to avoid the obstacles.

If you set it VFR ATC will let it fly it's own route and it will fly through hills etc.

 

I purposely set an AI aircraft to fly low in a hilly area and sure enough ATC told it climb, then when free of the hill it got told to fly low again.

 

This works for me anyhow.

 

 

Col.

Link to comment
Share on other sites

hgschnell,

 

The downloaded file name is: ugam_ambrolauri_georgia.zip and can be found at AVSIM and possibly elswhere as well.

 

This very light airfield did not exist within the basic FS9 but for that reason the author has included two approach BGL files, one for UGAM's RWY 11 and the other for RWY 29. However, these are good for user aircraft in which the pilot can dodge the mountains during the final part of his approach, but according to Jim's tutorial, these can also work for AI aircraft. Unfortunately though, AI aircraft do not dodge mountains during their approaches but fly straight through them and for that very specific reason Jim has made completely separate AI approach routes for some airports/airfields in mountainous areas, so that approaching AIs now visibly dodge the mountains in their very own AI way. It's this very specific Jim Vile AI approach technology on which I'm now busting my brains.

 

Hans

Link to comment
Share on other sites

Hans, I was involved in this project and we achieved AI approaches without flying through mountains.

 

It depends on the FDE of the AI aircraft.

UGAM is served in real by a LET L-410 of Vanilla Sky and this workk in FS9 (Do228 too).

 

The readme says:

"...

Known problems

--------------

Due to the mountainous area, FS9's AI traffic will have a somewhat erratic behaviour in spite of customized approaches by Hans-Günther Schnell.

Our tests have shown that it depends of every specific AI model, if it can use the approach and will land safely or if it flies through

mountains or even crashes at its way. It seems that the AI traffic engine of FS9 shows its limitations in this kind of environment.

..."

 

If you tell us what AI aircraft you used, we could have a look at the behaviour.

Link to comment
Share on other sites

Hans-Gunther,

 

Yes, my AI aircraft, which flies IFR between UGGG and UGAM, is a LET-410 but to my great surprise, the FDE (= Air file) specifies the Beechcraft KingAir 350 in it's description field and is therefore not the original AI LET-410 FDE. My AI LET-410 itself seems to fly quite normally.

 

However, I have 6 other seemingly very well working and non AI aircraft specific (I think Jim Vile's) AI approach files for the following airports/airfields: (Positioned in my Generic/scenery folder).

 

LOWI Innsbruck in Austria. 7Kb

LSGS Sion in Switzerland. 4Kb

LSZA Lugano in Switzerland. 3Kb

LSZL Locarno in Switzerland. 2Kb

LSZS Samedan in Switzerland. 1Kb

VHHX Kai Tak, in Hong Kong, which us still active in my FS9. 6Kb

 

I assume that the different file sizes represent their different degrees of difficulty.

 

These all seem to work correctly for AI approaches/landings by non-specific AI aircraft and I have tested the process at each airport/airfield by following different (test) AI aircraft from their take offs, then flying in formation with them and finally down to their correct landings. However, I must also admit that some of my heavier AIs could not exactly follow the sharper turns, e.g. to final at VHHX (Kai Tak) and would then land either next to the runway or in the adjacent bay. However, I was able to fix these problems by decreasing their empty weights and/or fuel loads.

 

Hans-Gunther, are you telling me that the two approach files supplied with the UGAM's download, should have the same function as the six I already mentioned above ? And if so, could the AI Beechcraft KingAir 350 FDE be the culprit ?

 

In the meantime I will search for a LET-410 base file and will do some testing with it's FDE file ...... if found.

 

Thanks for your help and stay healthy.

 

Hans

Link to comment
Share on other sites

Just out of interest one of the best air files in the default Boeing 737-400, I've used it many times for ai aircraft, if you look at CalClassics it used a lot on their aircraft.

 

I've had some flyable aircraft that work well as AI and some that don't and I've had to tweak them.

 

 

Col.

Link to comment
Share on other sites

Hans, your list could be completed with

 

LICR Reggio di Calabria, Italy and

KDCA Washington, DC.

 

These curved approachs allow a more or less intelligent landings. But they don't avoid that AI fly through montains some serveral miles before reaching landing paths. That especially happens at Innsbuck. Furthermore normal taking off from valley airports will remain very tricky or almost impossible. Even from LSGS Sion airport. In fact, creating AI flights to and from such airports doesn't make great sense.

 

Bernard

 

 

Link to comment
Share on other sites

Hans-Gunther,

 

I substituted my original AI LET-410 AIR file with one included in an AI LET-410 base files download and saw that it was basically a DH-6 Twin Otter AIR file. Did not notice much difference in the flight characteristics though but this (AI) Twin Otter at least externally looked a little more like an (AI) LET-410 than my originally used Beechcraft King Air's AIR file. However, I have the feeling that this in itself was not the definite "cure-all" for my original problems/questions.

 

I also increased the pattern altitude at UGAM to 2760 feet (ASL) because the original was even (far) below the actual airfield altitude of 1760 feet (ASL). However, I have not tested the effect this could have had any further.

 

I took off in one of my flyable LET-410s behind an AI LET-410 from Tblisi airport (UGGG) and followed it to UGAM while it was flying a VFR flight plan but it again disappeared into a mountain during the last stages of it's approach, seemingly to RWY 11. I then changed it's flight plan into IFR and again followed it.

 

Low and behold ....... this time I was able to follow the AI LET all the way through the valleys to a final touch down on UGAM's RWY 29 and to me this confirms that any AI flight plan involving UGAM as it's destination, MUST be IFR and not VFR, especially when a custom Jim Vile type AI approach is active. This then automatically also holds true for any other airports/airfields where Jim Vile type approaches are active. The only problem now left over is that AI take offs, cannot follow the same mountain dodging routes ...... yet.

 

My only criticism on this UGAM airport is that the README file is not very clear on this whole subject.

 

Anyway Hans-Gunther, thanks for the lesson and your part in the UGAM approach project, I have learnt a lot from it.

 

I also thank all you other guys for your reactions.

 

Regards

 

Hans.

Link to comment
Share on other sites

Yes Collin, you are right but without any mountains or other high ground in the direct vicinity of the destination fields, it doesn't make much difference, other than that approaching VFR AIs make quite some turns and twists before entering final, while IFR AIs make more straight in and ILS-like approaches.

 

For testing/checking purposes I quite often position my Grumman Tomcat somewhere on my airports/airfields and therefore also on UGAM, predominantly to check approaching and departing AIs.

Why a Grumman Tomcat ? Because it has a radar gauge in it's panel and on which I can follow airborne IFR and VFR AIs to/from maximum 40 Nms away. Not only that but I can also check their flight dynamics at different speeds.

 

As the result of your earlier IFR remark I checked all my AI flight plans involving the six airports/airfields in my list and changed all VFRs to IFRs. So, it's a big thanks to you as well.

 

Regards and happy landings.

 

Hans

Link to comment
Share on other sites

Guenther,

 

I have reacted to your private message but also have some initial technical questions for you, which I hope you can answer.

 

What are the basics for creating such a (mesh-)mountain dodging AI approach path all the way down a winding valley to a final landing at the valley's bottom ? E.g. What is the first point to which the AI plane is steered and what path(s) does it then follow to the airfield ? Is XML the only way of programming this ?

 

Do you know of any tutorial I can make use of ?

 

Fond regards

 

Hans

Link to comment
Share on other sites

Hans,

 

go back to post no. 2 of Tom Gibson, he had answered short, but very exactly to your question.

 

the tutorial is here:

 

http://www.fsdeveloper.com/forum/threads/tutorial-approaches.14074/

 

The trick is to force AI to an ILS approach without an existing ILS.

 

You can use BGL2XML to decompile some of Jim Vile's curved approaches (or mine) to understand the principles of operation.

Link to comment
Share on other sites

thanks for the information, I will give it a try.

 

I discovered Traffic View Board 2.1 installed on my PC long ago, but I did not use it much after I installed MS SDK Traffic toolbox.

 

I prefer it because I can see more internal details of an AI aircraft on the map.

 

Mostly is my focus on an approach and not on time tables (for time tables and parking I prefer AIFP)

Link to comment
Share on other sites

Hans,

 

go back to post no. 2 of Tom Gibson, he had answered short, but very exactly to your question.

 

the tutorial is here:

 

http://www.fsdeveloper.com/forum/threads/tutorial-approaches.14074/

 

The trick is to force AI to an ILS approach without an existing ILS.

 

You can use BGL2XML to decompile some of Jim Vile's curved approaches (or mine) to understand the principles of operation.

 

Here is another fine post:

https://www.fsdeveloper.com/forum/threads/curve-approach.16806/

Link to comment
Share on other sites

Guenther,

 

I have not tested UGAM any further because my AI approaches there are now all being performed correctly and even by different AI types. Very spectacular indeed and many thanks for all your work.

 

Normal piloted ATC vectored approaces also work correctly but during the last few miles the pilot really needs to know the way through the winding valleys. No problem though because this is what flying is all about.

 

However, my only remaining and very spectacular airport, where this AI auto-approach system does not seem to be able to work is VQPR, Paro in Bhutan, where the mesh mountains all around are so high and at such close proximities, that even flyable aircraft need very steep final approaches for landing.

AI take offs also remain a problem because, even after their related aircraft FDEs and CFG parameters have been edited, they cannot realistically climb steeply enough to avoid some mountain tops during their climb-outs.

 

In the meantime I've seen some YouTube movies about real Airbus 319 approaches/routes through the mountain valleys to both runway ends at Paro and to my great surprise, these can also be more or less followed correctly when approaching the FS9 depiction of this airport, but then again, once the aircraft has entered the initial valley, it's pilot really needs to know the way to the airport.

 

Thanks again for the lessons.

 

Regards

 

Hans

Link to comment
Share on other sites

Hi,

 

I've only ever used ADE to add approach code. Set it up to compile as "AFCAD Split" and it'll create a separate approach BGL file. I've also set it up to compile to a temp folder, so nothing gets overwritten by accident.

 

As said before:

 

One needs to set up the approach as an ILS, otherwise ATC/AI will only use that code in IMC conditions.

 

Only some of the approach header info in ADE is important for ATC/AI.

The fix (usually the FAF), that's when the AI will drop it's landing gear.

The alt. at which AI crosses that fix.

The missed approach alt.

The course the AI will fly after crossing the fix, in ADE in TRUE degrees. It will also use that course to to fly towards the fix at a ~30 degree intercept.

 

The old ADE 1.65 pdf can be downloaded here:

http://www.mediafire.com/download/8rp6jr7en9xwdls

 

 

Regards,

DDP.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...