Re: [Flightgear-devel] AI at the airport...
On Wednesday 23 April 2008 00:06, Christian Schmitt wrote: > > So does it mean that the "On the runway" points I set in taxidraw are > currently of no use? Current use is limited. I do use them to determine whether a taxi route crosses a runway. Currently, routes that cross a runway are given a penalty, so that the route-finding algorithm preferably returns an alternative route that doesn't involve runway crossings. As with other parts of the AI code, I'm planning to refine this, so that penalties are incurred only when aircraft are crossing active runways. FYI, I would expect it to be possible to set the "On the Runway" property automatically from within taxidraw within the next month or so, and will try to make use of that information inside of FlightGear within the next six months or so. > > Another issue I have is that setting a point as "Cat II/II" will lead to > a segmentation fault on next FG startup. > Okay, that seems like a regular bug. I'll have a look. Thanks for reporting. D. - This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] AI at the airport...
Durk Talsma wrote: > We are moving towards a more sophisticated runway exit strategy: Last year at > LinuxTag, Thomas Foerster and I discussed the idea of adding a performance > database that could be used to determine stopping distances, and I'm > currently working on adding support for runway "entrance/exit" points. > Ideally, I'd want taxidraw to set these automatically, but this is still on > my TODO list. So does it mean that the "On the runway" points I set in taxidraw are currently of no use? Another issue I have is that setting a point as "Cat II/II" will lead to a segmentation fault on next FG startup. Cheers Chris - This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] AI at the airport...
Hi Christian, On Tuesday 22 April 2008 20:42, Christian Schmitt wrote: > Hey Durk, > > Ralf just pointed me to you being the expert on AI/ATC and stuff like > that, which is IMHO one of the most important things for a good user > experience. :-) > As you might know we are currently doing a lot of work on EDDF. One of > the things I am doing right now is redesigning the AI-Network. AI Planes > are starting and landing on two runways. The runway where they start is > not a problem, but the landings are strange: The plane lands approx. in > the middle of the runway length and brakes. Now, in real life I would > suppose it to take the first taxiway available to leave the runway. > Well, it doesn't. Instead it continues on the runway till it reaches the > last possible exit, right at the end and takes it. > Is this supposed to be normal behaviour? This is currently normal behavior, but it is a limitation of the current code. Currently, FlightGear finds the node that is nearest the beginning (takeoff) of end (landing) of the runway, and creates a route from/to the parking position using these two points. We are moving towards a more sophisticated runway exit strategy: Last year at LinuxTag, Thomas Foerster and I discussed the idea of adding a performance database that could be used to determine stopping distances, and I'm currently working on adding support for runway "entrance/exit" points. Ideally, I'd want taxidraw to set these automatically, but this is still on my TODO list. Many of the bits and pieces are already in place, but the whole system needs more integration. I hope to address these issues, as I'm moving forward with integrating the AI / ATC code. > > Another question: Should I use a rwyuse.xml file? I have currently one > in use, but I don't like AI planes always taking the same end of the > runway for starting and not putting winds into consideration. Using a rwyprefs file will give you some more control over runway assignments, which might come in handy if you want to do more sophisticated stuff like parallel takeoffs / landings, or assign separate runways for general aviation / commercial and military traffic. If you want the runway choice to be more determined by the wind, you could try to play with the tailwind and crosswind parameters. In particular, setting these to smaller values would allow the wind to have a bigger influence. I realize that there isn't much documentation on the runway prefs file. have a look at one for EHAM, which is the most compliated one I have. If you have any questions regarding this file, please let me know and I'll try to hack together some documentation. Cheers, Durk - This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
[Flightgear-devel] AI at the airport...
Hey Durk, Ralf just pointed me to you being the expert on AI/ATC and stuff like that, which is IMHO one of the most important things for a good user experience. :-) As you might know we are currently doing a lot of work on EDDF. One of the things I am doing right now is redesigning the AI-Network. AI Planes are starting and landing on two runways. The runway where they start is not a problem, but the landings are strange: The plane lands approx. in the middle of the runway length and brakes. Now, in real life I would suppose it to take the first taxiway available to leave the runway. Well, it doesn't. Instead it continues on the runway till it reaches the last possible exit, right at the end and takes it. Is this supposed to be normal behaviour? Another question: Should I use a rwyuse.xml file? I have currently one in use, but I don't like AI planes always taking the same end of the runway for starting and not putting winds into consideration. Well I hope you can help me here. Thanks. Chris - This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel