February 8, 201412 yr I have AICull=3, yet AISIDSTAR is deleting aircraft at my monitored airports (most recently right now at SBFZ). Cheers, Bryn. Streaming at twitch.tv/brynmwr
February 8, 201412 yr I'm running the current 1.2Q on a network with simconnect and widefs configured on the client. When I launch aisidstar, the command prompt never changes from "Waiting to detect AI aircraft within 200000 meters, but says "FSUIPC 4 Link Established." I don't think it is working, what am I doing wrong? When AISIDSTAR starts, it provides a list of ICAO(s) that are being monitored. If you're at an ICAO that is not on that list (or within 100 nm of an ICAO on that list), then the program won't have anything to monitor. The program monitors an ICAO when it finds a valid SID or STAR file for the ICAO. I have AICull=3, yet AISIDSTAR is deleting aircraft at my monitored airports (most recently right now at SBFZ). Cheers, Bryn. Are you monitoring SBFZ? Does the ICAO show up on the list of ICAO(s) being monitored when AISIDSTAR starts? If not, then AICull (default off) is doing its job by deleting the AI at SBFZ. AICull is designed for flights between monitored ICAO(s). AI that is neither arriving at or departing from a monitored airport will be deleted, so that the user can have the experience of dense AI at the departure and arrival airports, but not the framerate overhead of AI at nearby airports (that the user won't be visiting). You can also add AICullHeight = 18000 to your .ini, which prevent AI above 18000 from getting deleted regardless of departure/destination. That way, the user will still see normal AI traffic densities at flight levels. -Roland
February 8, 201412 yr Has anyone set up an ini file for KPHX, KLAS? I'll play around with the military overhead approach, planned for Nellis AFB (KLSV), but that might be too complex for testing. I did but there are no approach files. I make approach files for airports that do not have straight in approaches.
February 8, 201412 yr AICull is designed for flights between monitored ICAO(s). AI that is neither arriving at or departing from a monitored airport will be deleted, so that the user can have the experience of dense AI at the departure and arrival airports, but not the framerate overhead of AI at nearby airports (that the user won't be visiting). You can also add AICullHeight = 18000 to your .ini, which prevent AI above 18000 from getting deleted regardless of departure/destination. That way, the user will still see normal AI traffic densities at flight levels.I really like the idea of this "Ai cull", why didn't anyone think of that? I was wondering if you can modify this feature a bit. Maybe some of us have a whole lot of airports monitored (I have AISIDSTARS monitor most of the major airports of the USA and Canada-not to mention the rest of the world. Its needed for the high traffic airports) but we may be bypassing some on the way to a different one. Eg KIAD to KEWR, I am not flying to DCA, BWI, LGA, or JFK but they are monitored. Can you create a file that the user specifies their departure, destination, alternates and or an (un)limited amount of airports to "save" from culling for that flight. This list should be progressively scanned so any changes while the program is running can be seen right away. With this, all AI will be culled via your already set parameters (ie 1,2 or 3) unless they are on the "safe list". If I overfly JFK which is monitored, I don't want to have traffic on the ground generated because it is not the airport I am going to. I hope you understand what I mean. thanks On another note: If you open traffic toolbox's map function, you see all the ai that is present in fsx's reality bubble. If you notice each flying aircraft has a node of sorts always in front of it. I think that node is the target "imaginary waypoint" that fsx ai engine uses to direct where the ai plane should go to. Once this node is arrived at, a new one is created and the process is repeated until the plane is landed at its destination. I noticed that the nodes contain altitude, speed, position( of course). These are the information "aiplayer" dll uses to know where to fly the plane. Now AISIDSTARS I believe does similar but allows preset nodes (waypoints) to be created. It would be a great experiment to see if you can interfere with what fsx is sending to that node. Suppose fsx commands speed 220 @ 13000ft somewhere, and according to the approach plate, the plane should be 180@ 9000ft here. Could it be possible to impose the approach plate's direction onto the fsx's? If this can be done, you wont here the atc repeating ignored commands as it stands now, but also the plane could be better controlled I think. Your thoughts... Cheers!
February 8, 201412 yr Thanks....I think that's possible....here's a link to some Toncontin final approach files I just wrote, probably the most "extreme" final approach file I've written yet. The same concepts could be adapted to military style approache Roland, great template and testing now with Nellis AFB, NV, but I'm sure there will be some .ini tweaking. I'll wait to update to 1.2Qbeta after testing.
February 9, 201412 yr Tested, SIDs work great for a busy mil base like Nellis, gets the AI jets on their way in an orderly fashion. I used both the KDCA and posted MHTG visual files as a template, but initally the AI were not being assigned. After figuring out my issues with the Final STAR's, I added "Jet" and that helped the AI assignment, although I still had a few not get assigned the STAR and just fly on. This also makes sence as generally prop cargo aircraft (C-130's) are not in the overhead, but this might look funny to the AI jet tankers/cargo models. The AI jets made the step down fixes at the commanded speed. The issues seems to be the AI flight dynamics, as the angle of bank is very limited, but probably fine for airliners. My HTAI F-16's could not make the turns at 300KIAS to line up for the overhead approach, normally going around to make the waypoint. When adding the points, I tried to be somewhat conservative with what the AI could do. I looked at changing the autopilot section in the aircraft.cfg file, but have read this is useless. I know FDE changes, even in AI aircraft, can be painful as one change causes trouble with 5 others.
February 9, 201412 yr New Rev. 1.2Q (beta) can be found here: http://www.mediafire.com/download/hc4vyl6hq1x1r0v/AISIDSTAR12Q.zip FIXED: AISIDSTAR sometimes crashes when writing final approach files to the autogen cache folder. Then, when user tries to restart AISIDSTAR, the program warns the user that one of the STAR files is formatted incorrectly (i.e., the program won't restart). Important Revision: A registered version (4.9x or later) of Peter Dowson's FSUIPC utility is recommended to use the full functionality of AISIDSTAR. For example, the new AICull function requires FSUIPC. For networked configuration users, Peter's WideFS (ver. 6.9) is also recommended. Please consult the following video documentation (Note: for Rev. M, but still mostly applicable to the latest revision). Video 1 - Working with the Included Sample SID/STAR Files and Creating New Files Using AIConv.exe: http://www.mediafire...sing_AIConv.mp4 Video 2 - Custom Editing SID/STAR Files - http://www.mediafire..._STAR_Files.mp4 Video 3 - Operating Parameters. Part 1: http://www.mediafire...ters_Part_1.mp4 Part 2: http://www.mediafire...ters_Part_2.mp4 Part 3: http://www.mediafire...ters_Part_3.mp4 Part 4: http://www.mediafire...ters_Part_3.mp4 Part 5: http://www.mediafire...ters_Part_3.mp4 -Roland
February 9, 201412 yr Because in this tryout phase it is necessary to observe a few AI´s to find stupid behaviours i have three more requests: Can you please make (or allow) the Monitor Window wider? I use a 1920 x1200 px screen for that Window. But to see the complete line it would be usefull to have it wider, because now, when i want to read the "Special Commands" i cannot read the full Pattern, when i shift the column of the Pattern, the Commands are out of sight... 2. Though it is possible to highlight one single Line, after the next Window update the highlightning disapears. Can you establish a function that the highlight of a flightnumber keeps alive once set? 3. I would appreciate filters very much. At big Airports there are dozens of sleeping AI´s which are frequent not too interesting to have them in the list, it would help a lot to switch them off case sensitive. Same e.g. enroute Traffic above a given Flightlevel. That all will help to keep the List cleaner. After all, i still don´t understand if there´s a stable system regarding turns of the AI´s. Example 1: AI flies down the STAR, followed by a handmade Transition. There is a waypoint 90° (say D001) left from the passed. The following is another WP 90° left, (say D002) and one more 90° left, the FAF. So the complete Pattern is a U-Turn, the distance is little wider than a standard 2 min. turn. -AI overfly D001, starts to turn left, not 90°, not 180° but something 110°, then straight. The heading is NOT to D002, so the AI passes D002 in a short distance. About 2 nm later the Aircraft stops, turns outright more than 150° with heading somewhere between D002 and the FAF. Adjusting AITurnScalar or DistMOE doesn´t help. Why does the Aicraft don´t catch the heading to D002? If it would do so it will be a perfect standard turn, ending up in a good allignement to FAF and Runway. Example 2. LOWI, handmade VFR Approach RWY 08. Same situation, but here the AI overfly D001 a few feet, turns right outright 90° heading D002 (obviously slewed), then the short way to D002 and again a slewed 90° turn. The reality shows, it is possible to fly that turn, so why the AI have to turn twice outright like a Helicopter? LOWI is generally an example for a missing function. Under no circumstances the AI received a Runway assignement after completing the STAR. So of course it starts circling close to the Airport. Right above LOWI he gets his Runway, yesterday the 26. Seconds later he gets a new pattern, Close STAR TUL3A (shortened). This would be great, you know, LOWI is demanding and usually approached over RTT. But instead of this, a few more seconds later the Pattern is switched to FINAL 08 (why?), of course that means, AI flies the complete way in opposite direction, turns outright 180° at RTT and then back the Final. If he would at least FLY a 180° turn, but allways that slewed behaviour... I would appreciate a function like NoFinalHeadingMoreThan90° and NoFinalDistance= though, so that it can´t happen to receive the Final Pattern above the Airport.
February 9, 201412 yr Dear Roland, As pomised a few weeks ago, I am reverting with some feedbacks on this promising programs. First on the SIDs, as most of us it seems, reading this forum, this part goes pretty smoothly. I had some problems with some AI plane on the ground, taxing to the runway and going straight out of the taxiway instead of making the turn towards the runway. I dont think your program controls AI on the ground yet (or is it some hidden code?) so it might be my AFCAD. A nice improvement, if feasible, would be turn when reaching a certain altitude (like at 3000'MSL, right turn to next waypoint) Second, on the Stars and approaches, where most of my concerns and issues still lie (in no specific order): 1) Would it be possible, if not already the case, to have an option to put AI on hold again for a limited number of times after release from a previous hold, if needed. I had the impressions that if the AI had hold before, it could be made to hold again at every waypoint after if the program deemed necessary. Would be interested to be able to disable that as it may not be too realistic. 2) SID or STAR altitude restrictions at or above or at or below or between (I note that you wrote in the readme it is not possible, but maybe someday it would be) 3) Option to disable final appraoch controlled by the program if no specific customized approach file was made, ie not use autogen approach but default FSX. (For now I tried to let your program focus only on the STAR part and disable all autogen approach, or I think I did by specifying 0 in the parameter). 4) I noticed that AI bank and remain in same bank position when acquired in slew mode by the program the first time. Some turns left are made with a bank to the right and vice versa, which is not too esthetic when following some AI. They tend to level the wings at some point but sometimes that weird bank attitude appears again. 5) AI climbing back up towards FL500 and above (I guess this is because I dont have a registered FSUPIC, but I thought that on go around they would follow fs default go around procedure, if set in the ini file) 6) Is it possible to input in the SID or STAR file different coordiantes format in the same file. I think it does work as the readme file says what are the accepted coordinqtes format, but just in case. 7) The most important to me, as I guess it is the main source of my issues: the distance from last waypoint of star to FSX IAF. How to efficiently place one or the other: 5nm in any directions, in the flying directions of the IAF? After exiting the star exit point, does the Ai directly head towards fsx IAF or to a certain heading like I seem to rember reading somewhere here? Could you elaborate a bit more on how the transition is made when passing back the AI planes to FSX after the star exit waypoint. On that regard, I had some plane completely disappear after this last waypoint. 8) Could you by the way elaborate a bit more on this 10 sec period of time:AIIAFDetectionDelay=10 (default) 9) 2 error code or note I had when running the program: error:24 sendid 101080 index 1 cbdata 24 reset altitude constant separation I hope this is useful to you and other users, I am constantly trying to improve my ini.file and SID STAR file for RPLL, some clarification on the above would be much useful. Thanks a lot! Best regards
February 9, 201412 yr I'm running the current 1.2Q on a network with simconnect and widefs configured on the client. When I launch aisidstar, the command prompt never changes from "Waiting to detect AI aircraft within 200000 meters, but says "FSUIPC 4 Link Established." I don't think it is working, what am I doing wrong? Here's more info from the log file. AISidStar is stuck on the cycle below. Writing Telemetry File. Beginning AI heartbeat checks.... Pattern Distance Vector Sort Invoked. Pattern Distances Updated Beginning Hold Calculations Ending Hold Calculations Anthony Cacciatore
February 10, 201412 yr I really like the idea of this "Ai cull", why didn't anyone think of that? I was wondering if you can modify this feature a bit. Maybe some of us have a whole lot of airports monitored (I have AISIDSTARS monitor most of the major airports of the USA and Canada-not to mention the rest of the world. Its needed for the high traffic airports) but we may be bypassing some on the way to a different one. Eg KIAD to KEWR, I am not flying to DCA, BWI, LGA, or JFK but they are monitored. Can you create a file that the user specifies their departure, destination, alternates and or an (un)limited amount of airports to "save" from culling for that flight. This list should be progressively scanned so any changes while the program is running can be seen right away. With this, all AI will be culled via your already set parameters (ie 1,2 or 3) unless they are on the "safe list". If I overfly JFK which is monitored, I don't want to have traffic on the ground generated because it is not the airport I am going to. I hope you understand what I mean. thanks On another note: If you open traffic toolbox's map function, you see all the ai that is present in fsx's reality bubble. If you notice each flying aircraft has a node of sorts always in front of it. I think that node is the target "imaginary waypoint" that fsx ai engine uses to direct where the ai plane should go to. Once this node is arrived at, a new one is created and the process is repeated until the plane is landed at its destination. I noticed that the nodes contain altitude, speed, position( of course). These are the information "aiplayer" dll uses to know where to fly the plane. Now AISIDSTARS I believe does similar but allows preset nodes (waypoints) to be created. It would be a great experiment to see if you can interfere with what fsx is sending to that node. Suppose fsx commands speed 220 @ 13000ft somewhere, and according to the approach plate, the plane should be 180@ 9000ft here. Could it be possible to impose the approach plate's direction onto the fsx's? If this can be done, you wont here the atc repeating ignored commands as it stands now, but also the plane could be better controlled I think. Your thoughts... Cheers! Thanks, actually the idea of AI Cull wasn't mine. A user suggested it to me because he correctly deduced that I was already processing the data to add this feature without a lot of code changes. It isn't perfect for all FSX configurations however. For example, UT2 keep re-injecting AI that is "culled" (for a limited period of time), which gets a bit annoying. I'll look into your suggestion. I'm sort of in a debugging mode now however. In the meantime, I would suggest quickly commenting out (e.g., "//") the airports and regions you're flying over at flight levels the .ini file. AI Cull might be a good candidate for a freeware stand-alone utility. If anyone out there wants to write it, please feel free. I can help by sending some C++ simconnect code snippets. I think you're right regarding the FSX waypoints. I think though I would have to hack it. I'm trying to stay within simconnect (although I broke down and utilized the FSUIPC erase function). -Roland
February 10, 201412 yr -AI overfly D001, starts to turn left, not 90°, not 180° but something 110°, then straight. The heading is NOT to D002, so the AI passes D002 in a short distance.About 2 nm later the Aircraft stops, turns outright more than 150° with heading somewhere between D002 and the FAF.Adjusting AITurnScalar or DistMOE doesn´t help.Why does the Aicraft don´t catch the heading to D002? If it would do so it will be a perfect standard turn, ending up in a good allignement to FAF and Runway. Could you try this file: http://www.mediafire.com/download/fyag4m6ggv6tuc9/AISIDSTAR12QTurnTest.zip Please let me know if the 180 deg. turns work better for you. -Roland
February 10, 201412 yr 1) Would it be possible, if not already the case, to have an option to put AI on hold again for a limited number of times after release from a previous hold, if needed. I had the impressions that if the AI had hold before, it could be made to hold again at every waypoint after if the program deemed necessary. Would be interested to be able to disable that as it may not be too realistic. All the hold modes do this (AIHoldingAllowed = 1, 2, 3, 4) except when the program is trying to maintain target landing rates. To eliminate this exception as well, add AIForcedReleasePriority =0 and AITargetLandingRate = 0 to your .ini. 3) Option to disable final appraoch controlled by the program if no specific customized approach file was made, ie not use autogen approach but default FSX. (For now I tried to let your program focus only on the STAR part and disable all autogen approach, or I think I did by specifying 0 in the parameter). Add AIFinalApproachAutoRecord =0 to your .ini. 4) I noticed that AI bank and remain in same bank position when acquired in slew mode by the program the first time. Some turns left are made with a bank to the right and vice versa, which is not too esthetic when following some AI. They tend to level the wings at some point but sometimes that weird bank attitude appears again. Yes, I do need to work on this. The bank is "aesthetic" in that the AI are getting slewed rather than really banking. However, the banking effect doesn't keep up with the AI during sharp turns. Eventually, I'll revisit this code. 5) AI climbing back up towards FL500 and above (I guess this is because I dont have a registered FSUPIC, but I thought that on go around they would follow fs default go around procedure, if set in the ini file) Yes, that is the program moving AI that have been forced to land without clearance (e.g., circling the airfield too long) to a far away location because otherwise they'll just sit on the runway and block things. I recommend FSUIPC because these problematic AI can actually be deleted from the sim. However, you seem to want to turn off autogen finals (AIFinalApproachAutoRecord = 0) in which case AI circling too long without a runway assignment will just vectored out of the area (no landings) in a normal fashion. 6) Is it possible to input in the SID or STAR file different coordiantes format in the same file. I think it does work as the readme file says what are the accepted coordinqtes format, but just in case. Yes! Yes, after exiting the star, the AI directly heads towards the FSX IAF. Regarding the transition between the STAR exit waypoint and IAF, just make sure there is room for the AI to make a standard turn to interecept the IAF. The program will let the AI try to intercept the IAF (under FSX control) in order to get landing runway clearance. If an autogen or custom final applies, the program will then land the AI. Otherwise, the AI stays under control of FSX for the landing. Regarding the AI disappearing after the last waypoint, that's probably the program deleting close pairs of AI. They're impossible to separate otherwise if they "pair up" on final. Add AIAntiClosePair = 0 to turn off. AIIAFDetectionDely gives working space between the start of final approach and the IAF (e.g., the plane time to turn toward final in a normal manner before the high precision final approach begins, which would cause the AI to turn very quickly onto final). Regarding the warning messages, annoying, but you should be ok. -Roland
February 10, 201412 yr Here's more info from the log file. AISidStar is stuck on the cycle below. Anthony, when AISIDSTAR starts, are you seeing the ICAO you're at (or near) appear in the list of ICAO(s) that are being monitored? If not, that means the program hasn't read any SID or STAR files for that ICAO, so there is nothing to monitor. Double check your paths in the .ini file, where you have to point the program to a particular directory you want to load. -Greg
February 10, 201412 yr Hi Roland, i´m sorry to say, the new rev. don´t work for me at all. Seems it is worse... I try to describe it in Order: To stay focused at only few AI´s i only tinkered with the STAR BIKM1X and a following transition. Behaviour 1. CONDOR1234 is comming from south, passing BIKMU and get RONAD as next. Passing RONAD, next WP DL530, which is the first WP of the "Final" File. Instead of turning left for about 100° the AI is turning right and doing a preety big circle with 260°, heading DL530. Why? The Monitor says, its a Holding because lost of serperation conflict with Aircraft ID=1 Scrollig back in the Terminal Window and i couldn´t belive it: The AI with the "ID 1" IS Condor 1234! It "saw himself"... Restart Behaviour 2 CONDOR1234 is comming from south, passing BIKMU and get RONAD as next WP. Passing RONAD, next WP DL530. AGAIN a holding over RONAD, this time because of a Eurowings Aircraft less then 3 Miles. I tried hard to find it, finally i located it in the area between VALSU and NIKOG, (not at the picture anymore), thats about 10-12 nM! Restart Behaviour3 Lufthansa1111 is comming from south, passing BIKMU and get RONAD as next WP. In the meantime Air Malta is comming close from the right, 58kts fast. From behind in his trails a Eurowings with 240 kts. They are all joining together and i had have a package of 3 AI´s with 58 kts snailing trough the air. Finally the Eurowings won the rumble and accelerate, heading DL530. Right over DL530 the AI slow down to 58kts and turns outright like a heli. Same at DL550. But in the meantime one more Aircraft came along over another north STAR, passed RONAD and followed exactly the drawed line at DL530. WOW But before i got lucky i saw first a assigned WP DL550, then the Aircraft turning right (?), then i missed the AI in the list and finally i found it "enroute" controlled by FSX *wallbash* So i tried it several times, but it tends to be impossible to get a single aircraft on the track because... whatever. The tool struggle to decide which Aircraft is in front and which has to slow down, the tool struggle to keep or get control over the Aircraft. One Aircraft was circeling 30 minutes in the LOC area with assigned RWY 5R but did not get Patterns. The tool struggle to control which "diameter" a turn should have. And finally the new rev. seems to assign almost each turn in the wrong direction. So, in my opinion it is currently impossible to get valid results in my "testbed" with UT2. Basically the only way is going back to the roots, simple Airport, only 1-3 Aircrafts and then checking the basics, mathematics, rules. Nothing for a public beta, rather doing homework.
Create an account or sign in to comment