FSX-MS Ai Controller (Sid, Jet Route, Star And Final Approach) Beta

[tf51d 284]

For almost perfect spacing look into FS AI Separation. Desired should be 4-5. Minimum about 2-3. Area about 35nm.

 

I'm talking about at cruise altitude, not approaches, correct me if I'm wrong, FS AI Separation (like AISmooth) only works during the approach and landing phase?

[Insurrection_and_Stupidity 9]

 

 

Just to verify, are you using the UT2Compatibility = 1 parameter? If you enable that parameter, you might also want to bump AIMaxSpawnDistanceFromSTAR down around 70.

 

Checked - Yes - UT2 Compatablilty = 1

Checked - Yes - AIMaxSpawnDistanceFromSTAR = 70


In UT2Compatibility mode (UT2Compatibility = 1), the AI should not turn around back to the airport after existing the SID, instead the AI will either continue on a great circle to the destination (AIJetRoutesEnabled = 0), but will not join a UT2 jetroute, or it will join an AIController generated jetroute (AIJetRoutesEnabled = 1). Edit: I just realized this contradicts what I told you earlier (AI will pick up UT2 route), sorry. Do you usually follow the AI when it departs? If not, as your reality bubble moves (e.g., your user aircraft is enroute), UT2 should inject AI into your reality bubble as normal on assigned (or custom) jetroutes. AIController won't interfere with UT2 for AI farther than AIMaxSpawnDistanceFromSTAR (for arrivals) and AIMaxSpawnDistanceFromSID (for departures) from a monitored airport (assuming AIJetRoutesEnabled = 0).

 

Checked - Yes - UT2 Compatablilty = 1

Checked - Yes - AIJetRoutesEnabled = 0

 

OK - I am not actually flying, but sitting stationary at EGPF Gate 26 so my reality bubble is not moving.  What i was expecting was that once an aircraft leaves the SID, which is being flown perfectly BTW, I was hoping that it would then pick up the next waypoint in the UT2 generated flightplan and continue on its way.  If that was possible then this would be awesome.  I shall download V1.3b and continue my obs.

 

I also noticed that if I enable AITrafficPattern = 0, once an aircraft lands at the end of the rollout it will stop then disappear.  If disable this then landing and rollout is normal with subsequent taxi to the gate.

 

Thanks Roland.

 

Best

 

Ken

[jabloomf1230 2,445]

 

 

 BTW, does anyone know if there is a way for the thread start to update the original message?

 

Try creating a  new post that has the most recent download link. Then see if you can "Mark Solved" that post making it the "best answer". You should be able to keep doing that as the thread originator. I'm not sure that it will work, but it's worth a try. Maybe one of the Mods here can explain it better than I did.

[russianspd 10]

I'm talking about at cruise altitude, not approaches, correct me if I'm wrong, FS AI Separation (like AISmooth) only works during the approach and landing phase?

Nope, it separates any airborne traffic. You could also choose to separate ground traffic but that would mess up AI taxiing and would cause slews. AISmooth's approach to separating AI is to put them into holds, having this done at any airport with more than 4-5 aircraft in the approch area is high unrealistic. With AI Separation, it slews the AI the value of the minimum nm distance if the aircraft falls below the desired nm value of the AI it's following. Thus, you get AI perfectly spaced on the arrival corridor making landings much more fluid. 100% AI at LAS with landings on both 25L and 19R average me 10+ arrivals every 10 mins or so. About a landing a minute or so. Very useful. Just got to make sure you AFCADS are configured right to have AI be able to vacate the runway properly.

 

I found using AISIDSTAR that even though I had separation active and set that AI still bunched up or came in too close to one aother. Using AI Separation along with AISIDSTAR when I did testing at LAX gave me roughly 1 go around for every 5+ landings. No holds or anything.

[ysfsim 22]

 

 

Yes, it will be aware of all aircraft position data (including occupied gates and parking areas) and whether any AI or user aircraft is getting too close to AI controlled by AIController. The general plan is that if an AIController controlled "AI" comes into conflict with an AI controlled by some other program (e.g., AI pointed nose-to-nose on the same taxiway is a good example provided by Jim) or user aircraft, the AIController "AI" will be deleted to avoid problems and maintain compatibility.

Can't it just reroute "your" ai around such conflicts when able? Is it too much to suggest a user file for particular airports that have special or specific taxiway usages? ex. KMIA, Lufthansa's A380 can only taxi via certain taxiways. can we have a file that we can edit to specify parameters like that?

 

Your progress sounds exciting, cant wait for release!

[funkyj4ever 2]

Still it doesnt work... I have a complete new computer only have the same gpu and ram. Also a complete new windows install. Still the aircraft freeze in a strange position. 

 

http://puu.sh/duMoJ/2efaaafc34.jpg

 

Like on this pic. It flies just like this in a straight line. I use ice traffic. Do others have ice traffic and have it working? 

[jabloomf1230 2,445]

 

I use ice traffic. Do others have ice traffic and have it working? 

 

 

You might want to refer to this thread. I doubt anyone here will help you.

 

http://forum.avsim.net/topic/432570-ice-ai-traffic-group/

[grandfred29 180]

My traffic fly with the same position as your plane when aijetroute=1

Change to 0 and it Will be good

[jasoncardeira 30]

Hello guys, i started to read this today, and this seems a very good and useful tool! Obviously i will install this but i need to know some things:

 

1 - is it necessary to have fsx sdk installed? I have installed fsx sp2 only.

 

2 - will this work with mytraffic 2013?

 

3 - do i have to download the v1.2 on avsim and apply the v1.3 after that?

 

Thanks

[RancidViper 37]

 

 

I tried this and I sat plane spotting but didn't see a single arrival at EGCC for a whole day (12hours) of testing, SID's work perfectly but STAR's just never seem to happen. I see the arriving traffic being sent to the beginning of the STARs, but every one of them disappears from FSX before it manages to get to the STAR start point. They all seem to climb from their initial spawn points of between 7,000 and 15,000 feet and then climb to somewhere in the upper 20,000 to 30,000 feet or more and sometime after that they just disappear completely.
 
I've tried this with both MTX and UT2 and various spawn distances and had no joy.

 

I have this same issue, but at LOWI. Started to follow aircraft that I knew were supposed to come in for landing, but they just keep flying way off course until they just vanish completely. Didn't see a single plane coming into land. The ones that take off seemed ok, just that they just kept flying east through the valley, not sure if that's the correct procedure, but it's still better than it was, which was a sharp left turn straight into a mountain.

 

Any ideas what the issue is? I'm using MTX.

[BuboBubo 35]

Hello guys, i started to read this today, and this seems a very good and useful tool! Obviously i will install this but i need to know some things:

 

1 - is it necessary to have fsx sdk installed? I have installed fsx sp2 only.

 

2 - will this work with mytraffic 2013?

 

3 - do i have to download the v1.2 on avsim and apply the v1.3 after that?

 

Thanks

 

 

1. According to the read me it requires "FSX SimConnect SDK" and "For full functionality,

AIController requires the latest, registered version of FSUIPC ver. 4.9x or later"

2. yes.

3. No, I downloaded and installed version 1.3A directly without installing any previous versions.

[jasoncardeira 30]

Ok, thanks for your answer

[FlyTweety 2]

Without wanting to push to hard Roland, but do you think, we see the "non IAF" Update during the X-Mas holidays? :ph34r:

[Clipper Ocean Spray 189]

Try creating a  new post that has the most recent download link. Then see if you can "Mark Solved" that post making it the "best answer". You should be able to keep doing that as the thread originator. I'm not sure that it will work, but it's worth a try. Maybe one of the Mods here can explain it better than I did.

 

Thanks, appreciate it.  I might give it a try when I upload the next version.

 

 

I have this same issue, but at LOWI. Started to follow aircraft that I knew were supposed to come in for landing, but they just keep flying way off course until they just vanish completely. Didn't see a single plane coming into land. The ones that take off seemed ok, just that they just kept flying east through the valley, not sure if that's the correct procedure, but it's still better than it was, which was a sharp left turn straight into a mountain.

 

 

Ugh, LOWI should work, there's custom STAR and final approach files for it.  A readme.pdf links to a demo video showing the approach being performed.  The AI does go out of the way a bit by intercepting RATT (Rattsburg?), turning, then descending through the valley.  Does the AI even seem to be intercepting RATT?

 

Anyway, the whole approach section is being massively re-worked right now.  

 

Without wanting to push to hard Roland, but do you think, we see the "non IAF" Update during the X-Mas holidays?

 

 

I don't know, I would think probably about mid January (the proof-of-concept issues are behind me).  

 

I'm working on crossing restrictions right now (e.g., AI doesn't taxi across runway 27L when another AI is on short final or beginning a departure).  The next item is detecting conflicts on the taxiway.  I also need to revise the AIConverter to read final approach nav data.  The last item is working through the list of bugs/suggested improvement for versions 1.3A and B (in regard to a substantially different version 1.4).  

 

I'm also thinking about taking over the takeoff rolls too (relatively easy to do) because personally for me one of the most annoying aspects of FSX/P3D is how the AI slowly lines up on the runway,then STOPS for about 10 seconds!

 

Looking back, the decompiler took the most development time (although it was very interesting work).  Ironically, the decompiler will do its work behind-the-scenes in the new version.

 

-Roland

[SierraHotel 979]

Hi Roland, thank you so much for all your hard work.