After the night my memory starts to work again. There happend so much things at the same time, it was overwhelming :rolleyes:

 

Regarding the 2 Aircrafts flying back and forth: It happend at the Base Distance Leg. I wanted to watch the turn more close because in my opinion the turns are still to sharp, Therefore i slewed with my "Aircraft" BOB to the Base Entry, not aware that the minimum separation between AI and User Aircraft was set back on 7 miles (i noticed, there´s a 1.5 miles value now wen i set all to default).

However, for the first AI that meant that is was send back to the Base Entry point (had almost arrived the final turn point), then it got under the minimum separation  with the following AI at the Base Leg, this AI was also sent to the Base Entry, and because my Base Leg is shorter than the minimum Separation (5 miles vs. 7 miles) it was a self holding situation also as i left the area.

 

A good question is if i had 2 instances: I would say 95 % NO, but to be honest, there´s a rest risk of 5 % that there have been a instance minimized?

However, in the future i will save a copy of my ini before i tweak anything, and of course before i update AIC.

Just loaded HotFix 6 and the 'doggie style' parking is back. 

 

I was at YPPH and gate 13 had an AI on stand and it had been there for quite a while as prepared for departure from gate 14. An arriving 767 was given taxi instructions to gate 13 and when it arrived it remained for a few seconds then disappeared. is this the way it is meant top work?

Hi, I searched this thread but I couldn't find any experiences in terms of performance.

 

The reason for my search is the following:

When I use AIController, I have huge stutters after some time (once per second). It seems to me that AIController takes a lot of CPU power to manage the full AI environment.

It wouldn't surprise me as I've set my traffic at 100%. All airports worldwide (respectively all converted SIDSTAR files) are active at the same time.

But I also have to mention that I use AirTrafficManager, it limits the traffic at 100 aircrafts.

First thing I would try is to turn off the handling of ground traffic. The earlier versions without this feature, eg V1.3, worked well, even at 100% traffic.

 

I'm wondering which of the features (jetroutes, takeoff/approach handling/ground handling) has the greatest performance impact generally?

Any ideas and/or own experiences are highly appreciated! Thanks!

Just loaded HotFix 6 and the 'doggie style' parking is back. 

 

I was at YPPH and gate 13 had an AI on stand and it had been there for quite a while as prepared for departure from gate 14. An arriving 767 was given taxi instructions to gate 13 and when it arrived it remained for a few seconds then disappeared. is this the way it is meant top work?

Further to the above post, just left YPPH and while taxiing two flights arrived both Triple7's - one SIA and the other UAE - and both were directed to an already occupied Gate 55 (the gate is identified for both airlines). These instructions were given even though Gates 52 and 53 were unoccupied both being identical in size to Gate 55 and neither having any specific airlines identified. Should AiController have chosen one of the vacant gates?

Hi Roland,

 

 

 

NEW:  Support for coordinated STAR-transition-approaches added (i.e., preference is given to selecting an approach having the same transition waypoint as the STAR exit waypoint the AI was assigned to). 

 

Thanks for the update, and your relentless developent of this outstanding little programm.

 

Unfortunately it seems HF6 did not eliminate the problem with the wrong transitions being selected. At LOWW - all traffic is still receiving the BALAD transition regardless from which star they arrive...

 

Did I miss a certain UI setting or are others also experiencing this little flaw?

 

as you can see on the attached Picture: Tyrolean803 and TURKAIR1886 came down the NATEX2W STAR and at WP PESAT they schould follow the PESAT transition, but are diverted to the BALAD transition....

 

https://drive.google.com/file/d/0BwD0Mexs_kFCVHd3eDB3Z3l0N2s/view?usp=sharing

 

 

kind regards

 

Richi

Edited April 20, 201511 yr by liketofly1

Hi Roland,

 

One more little (tiny little) thing I observed watching the traffic via FlightSim Commander... AI Traffic tends to read waypoints in a way that they fly over - before they initiate the turns... now with shallow turns thats not really a problem.... with steep turns they do overshoot quite. Would there be a way to regulate this overshoot except with speed scalars? Can we make them initiate the turn a little earlier? Like for 45° turns....1nm 90° turns 2 - 2,5nm and over 90° turns 3 - 3,5nm earlier? (mybe my values would need a little adjustment)

 

OK - I know thats complaining at a very hi level - anyhow - just an Idea...

 

kind regards

Richi

 

P.S.: SID's work great!!!

OK, i´m a lucky Dude because i´m at home tonight.

 

So i have the time to test again and guess what? It happend again.

 

I was able to make a screenshot in FSC.

 

 

The situation is starting when the WIZZAIR arrived the blue dot and ORENBURG is at the red dot. OF COURSE an idiot is lining up at the runway :ph34r:

 

There i get the first Message:

 

 

At the same time WIZZAIR turns back to the Base Entry WP.

 

In the meantime ORENBURG (S7) comes closer and finally they are both below minimum separation but AIC says:

 

 

However, it ends up in the situation of the first picture, both are trapped between Base Entry and Start Final Lineup WP, flying forth and back until they are out of fuel...

And of course, the Air Berlin was trapped too a minute later :rolleyes:

 

So, obviously there is an issue inside the Enhanced Autogen Approach that is running wired when one AI is performing a Go Arround...

When I use AIController, I have huge stutters after some time (once per second). It seems to me that AIController takes a lot of CPU power to manage the full AI environment.

It wouldn't surprise me as I've set my traffic at 100%. All airports worldwide (respectively all converted SIDSTAR files) are active at the same time.

But I also have to mention that I use AirTrafficManager, it limits the traffic at 100 aircrafts.

 

Yes, it looks like you have AIController maxed out on the local machine in terms of AI percentage and all airports with navdata being monitored worldwide.

 

I would seriously look into moving the program to a networked client machine.  AIController was built from the beginning to support a networked installation.  Even an older laptop (preferably running Win7) should have no problem running the program.

 

 

First thing I would try is to turn off the handling of ground traffic. The earlier versions without this feature, eg V1.3, worked well, even at 100% traffic.

 

Yes, just turn off enhanced mode, which turns version 1.4B into essentially version 1.3B.  Or...you can just stick with 1.3B.

 

 

I'm wondering which of the features (jetroutes, takeoff/approach handling/ground handling) has the greatest performance impact generally?

 

Probably writing the telemetry file (in terms of disk I/O).  It can be avoided by running AIController by itself with the telemetry option off, however you won't be able to use AIMonitor.  Also moving the program to a networked machine would eliminate these issues.

 

 

Further to the above post, just left YPPH and while taxiing two flights arrived both Triple7's - one SIA and the other UAE - and both were directed to an already occupied Gate 55 (the gate is identified for both airlines).

 

I'm having problems reproducing this, but I"ll keep trying.  I briefly had the wrong version of hot fix 6 in the download link, can you verify the AIController.exe version is 1.4.1.5?

 

The deletion while parked at gate is only intended if an AI is injected by a traffic program (e.g., UT2) into a gate already occupied by a parked AI.  In such a case, the parked AI is deleted to make way for the injected AI, so that the new AI can depart.  It isn't intended to delete an AI if two AI are assigned the same gate after landing.  If that is what happening, that is indeed a bug.

 

 

Unfortunately it seems HF6 did not eliminate the problem with the wrong transitions being selected. At LOWW - all traffic is still receiving the BALAD transition regardless from which star they arrive...

 

Did I miss a certain UI setting or are others also experiencing this little flaw?

 

Found this bug.  If you can believe it, it was all attributable to trailing whitespace during a string match.  Thanks for letting me know.

 

FlyTweety, this may also be the problem you're experiencing as well.  I'm not 100% sure, but I think I we should try the fix to see if it eliminates your problem.  I'll make the fix available shortly.

chumley, on 20 Apr 2015 - 9:06 PM, said:

 

Further to the above post, just left YPPH and while taxiing two flights arrived both Triple7's - one SIA and the other UAE - and both were directed to an already occupied Gate 55 (the gate is identified for both airlines).

 

I'm having problems reproducing this, but I"ll keep trying.  I briefly had the wrong version of hot fix 6 in the download link, can you verify the AIController.exe version is 1.4.1.5?

 

The deletion while parked at gate is only intended if an AI is injected by a traffic program (e.g., UT2) into a gate already occupied by a parked AI.  In such a case, the parked AI is deleted to make way for the injected AI, so that the new AI can depart.  It isn't intended to delete an AI if two AI are assigned the same gate after landing.  If that is what happening, that is indeed a bug.

 

Sorry Roland, I wasn't as clear as I might have been. My apologies if you have been searching the 'haystack' for a non-existent needle.

 

I am running 1.4.1.5.

 

Gate 55 was occupied. 

 

The UAE flight was assigned to Gate 55 and taxied to it stopping over the top of the 'sleeping' AI. One of the AI disappeared (not sure which as I was taxiing to the active and observing airport activity on EFB. While I was still taxiing the SIA AI arrived and was assigned Gate 55 by AiController. I didn't see the outcome as I had taken off.

 

What I am interested in is if all gates identified for SIA are taken should AiController direct any SIA AI to a vacant but compatible gate?

 

This has given me the idea it would be nice to see the gate number shown against the 'sleeping' AI if possible and also in the column for aircraft type i.e. Jet, Turbo etc. could you display the aircraft type as A330, B737 etc.?

Hi Roland,

 

 

 

Found this bug.  If you can believe it, it was all attributable to trailing whitespace during a string match.  Thanks for letting me know.

 

Super, excellent work! Thanks for your support..

One more little (tiny little) thing I observed watching the traffic via FlightSim Commander... AI Traffic tends to read waypoints in a way that they fly over - before they initiate the turns... now with shallow turns thats not really a problem.... with steep turns they do overshoot quite. Would there be a way to regulate this overshoot except with speed scalars? Can we make them initiate the turn a little earlier? Like for 45° turns....1nm 90° turns 2 - 2,5nm and over 90° turns 3 - 3,5nm earlier? (mybe my values would need a little adjustment)

 

I use turn-in calculations for taxi and final approach, but I was just looking at the taxi code and there was a bug (calculating turn bearing to the "waypoint + 1" using two latitude values!), so thanks for getting me to look at this part of the code again.  I'll look into adding it to the STAR/SID AI movements, the reason I haven't done so yet is the decreased need for turn precision during those sections of flight.

 

 

So, obviously there is an issue inside the Enhanced Autogen Approach that is running wired when one AI is performing a Go Arround...

 

See my post to Richard regarding a runway string-match bug, this could be the cause for your  issue too (worth a try).

 

 

The UAE flight was assigned to Gate 55 and taxied to it stopping over the top of the 'sleeping' AI. One of the AI disappeared (not sure which as I was taxiing to the active and observing airport activity on EFB. While I was still taxiing the SIA AI arrived and was assigned Gate 55 by AiController. I didn't see the outcome as I had taken off.

 

Ok, so it is possible an AI was injected to the gate (in sleeping mode) and the arriving AI already parked there was delete?

 

 

What I am interested in is if all gates identified for SIA are taken should AiController direct any SIA AI to a vacant but compatible gate?

 

Yes, it should do that.

 

 

One more little (tiny little) thing I observed watching the traffic via FlightSim Commander... AI Traffic tends to read waypoints in a way that they fly over - before they initiate the turns... now with shallow turns thats not really a problem.... with steep turns they do overshoot quite. Would there be a way to regulate this overshoot except with speed scalars? Can we make them initiate the turn a little earlier? Like for 45° turns....1nm 90° turns 2 - 2,5nm and over 90° turns 3 - 3,5nm earlier? (mybe my values would need a little adjustment)

 

Wow...thanks again..I'm looking at the revised code now and this bug fix is causing a massive improvement in turns while the AI is taxiing (e.g., the AI is "rolling" through turns much more consistently).  Edit:  it is really starting to look like the FSX taxi now...except I'm controlling the AI at just one-sample-per-second via an external channel)!  B)

Wow,looks like a promising weekend in case you release all improvements until friday :-)

Roland...it's awesome you have been on top of this program, considering this is not netting anything for you other than self gratification (and ours of course)!

 

I also want to throw out a thanks to those of you who are testing this for Roland!  I wish I had more time to devote to help out (I've piped in a few times), but some of you are doing awesome work!!  I'd like to take a second and thank you all!

Hi Roland,

 

do you think you would be able to manipulate the nav light control? In FSX is bit unrealistic and don't know why the logic was changed in FSX from FS9. In reality airliners switch on nav lights on the terminal some minutes before departure during pre departure preparation. Then the red beacon lights shortly before engine start, then the taxi lights, then the strobe lights when entering the take-off RWY together with the landing lights. Well, don't know how far you can go Just a thought!

 

Brgds,

 

Axel

Hi Roland,

 

 

 

Wow...thanks again..I'm looking at the revised code now and this bug fix is causing a massive improvement in turns while the AI is taxiing (e.g., the AI is "rolling" through turns much more consistently).  Edit:  it is really starting to look like the FSX taxi now...except I'm controlling the AI at just one-sample-per-second via an external channel)! 

 

Always there to help - you can thank me with a paypal donation button next to your awsome downloads... I'll happily support this development.... and I'm pretty sure others would also ;-))

 

regards