Board index FlightGear Development AI Traffic

Populate AI Traffic with real traffic

Intelligent, computer controlled vehicles that drive/fly over the planet!

Re: Populate AI Traffic with real traffic

Postby Hooray » Thu Nov 01, 2012 5:08 pm

Johan G wrote in Thu Nov 01, 2012 4:51 pm:Maybe time to make the pilot list scrollable? :wink:



Or maybe time to get a bigger screen that has no problem showing a list of 200+ aircraft ? :D
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11738
Joined: Tue Mar 25, 2008 8:40 am

Re: Populate AI Traffic with real traffic

Postby Johan G » Fri Nov 02, 2012 11:26 am

Hooray wrote in Thu Nov 01, 2012 5:08 pm:...a bigger screen that has no problem showing a list of 200+ aircraft ? :D

In my dreams, Hooray, in my dreams. :lol:
Low-level flying — It's all fun and games till someone looses an engine. (Paraphrased from a YouTube video)
Improving the Dassault Mirage F1 (Wiki, Forum, GitLab. Work in slow progress)
Johan G
Moderator
 
Posts: 5691
Joined: Fri Aug 06, 2010 5:33 pm
Location: Sweden
Callsign: SE-JG
IRC name: Johan_G
Version: 3.0.0
OS: Windows 7, 32 bit

Re: Populate AI Traffic with real traffic

Postby Michat » Fri Nov 02, 2012 9:41 pm

I thinks is a very good tool for users around the world, whose are conceited with the real chem trails commercial and private air traffic that populates our skies today.

Could be a good tool for detecting and denounce this kind of nasty activities.

Anyway is good tool for FG, so congratulation, your work is excellent and very appreciated.

Michat
User avatar
Michat
 
Posts: 1035
Joined: Mon Jan 25, 2010 6:24 pm
Location: Spain
Version: 191b
OS: GNewSense

Re: Populate AI Traffic with real traffic

Postby Hooray » Fri Nov 02, 2012 11:13 pm

well, all that info is freely available on websites like http://www.planefinder.net - so FG is not intended to suddenly become a tool for such purposes, there are better solutions for this fortunately.

Also, while we have a nice "demo" now, there are still a handful of more or less serious issues to really come up with a compelling prototype that's ready for "primetime".

This includes things like segfaults (crashes) and memory issues (leaks) that still need to be teased out, but also fixes to the "extrapolator", which still has some bugs and shows weird movements/orientation at the moment.

Also, there's currently talks going on to generalize the current design such that it's no longer provider-specific (planefinder.net), and also not specific to supporting just aircraft.
In fact, ship AIS has been mentioned previously, and it would obviously be another use case here.

And then again, there are some real technical limitations on the fgfs/fgms side of things, especially with regards to the existing inefficient MP protocol, which simply isn't suitable to be used for updating 200+ entities regularly.

To make really promising progress here, we would need to add another "output adapter" that doesn't use the fgms network/protocol but instead the existing HLA/OpenRTI infrastructure to send out derivative motion information instead of position updates, but that will definitely take another 1-2 FlightGear releases (at least) ...

Also, let's keep in mind that there's a more holistic approach being discussed for the whole AI traffic system, to factor it out into its own standalone component that uses the MP system, like Durk and the fgms developers mentioned several times:

viewtopic.php?f=23&t=6807&p=60580&hilit=#p60649
http://www.mail-archive.com/flightgear- ... 13222.html
http://www.mail-archive.com/flightgear- ... 13223.html
http://www.mail-archive.com/flightgear- ... 13245.html
http://www.mail-archive.com/flightgear- ... 31984.html
http://www.mail-archive.com/flightgear- ... 16400.html
http://www.mail-archive.com/flightgear- ... 13210.html
http://www.mail-archive.com/flightgear- ... 13221.html


All taken from: http://wiki.flightgear.org/Decoupling_t ... fic_System
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11738
Joined: Tue Mar 25, 2008 8:40 am

Re: Populate AI Traffic with real traffic

Postby adrian » Sat Nov 03, 2012 8:33 am

Hi,
I haven't read the whole thread yet, I just got a PM after being gone a while. Just my quick observations: our current AI aircraft system allows the ATC and AI pilots to react and interact between them and with the human pilot, inclucing via radio communication. This system would probably not. Regarding AIS ships, I had a world static snapshot which did not infringe on any terms and could populate (statically and only for visual purposes) harbors and seas. I might even have written some C++ code myself, but there's a problem: there are still lots of areas with low res terrrain where harbors and rivers are not well defined and the net effect is bad.

However, your ideas are quite nice, and a worthy hack, congratulations. I will revisit this thread, time allowing, and add some more comments.
Cheers,
adrian
 
Posts: 362
Joined: Wed Sep 15, 2010 2:15 pm

Re: Populate AI Traffic with real traffic

Postby Hooray » Sat Nov 03, 2012 1:55 pm

adrian wrote in Sat Nov 03, 2012 8:33 am:Just my quick observations: our current AI aircraft system allows the ATC and AI pilots to react and interact between them and with the human pilot, inclucing via radio communication. This system would probably not.


That's absolutely true obviously. And it would be a lot of work to change that, probably by using Philopsopher's idea and merely using the downloaded traffic data to populate our real AI traffic system to create interactive AI flights.
For the time being, the idea is just to enrich the MP experience and actually show real flights - taxiing/starting/landing are also not modeled properly. But if you like the idea of tracking a real flight via FG, then this is for you.
Especially, keeping in mind the low number of MP users we have.
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11738
Joined: Tue Mar 25, 2008 8:40 am

Re: Populate AI Traffic with real traffic

Postby F-JJTH » Sat Nov 17, 2012 8:26 pm

For those who are interested by this topic : currently FGAIS works fine when only 1 client is connected. We are able to see real live traffic in FG.
The problem is that FGAIS is not able to manage more than 1 client at same time ( I have a segfault with more than 1 client )

If a C++ programmer is able to spend some of his time in order to find how to fix this issue it's cool :)
The source code is available here : https://gitorious.org/~f-jjth/fgms/f-jjths-fgms-0-x

Cheers,
Clément
User avatar
F-JJTH
 
Posts: 697
Joined: Fri Sep 09, 2011 11:02 am

Re: Populate AI Traffic with real traffic

Postby Hooray » Sat Nov 17, 2012 8:45 pm

I know what the problem is, and I know that there are quite some other bugs. However, when I checked your wiki updates, you made lots of additions saying that those problems were solved already, so I assumed that you were using a locally modified version?
Like I said a while ago, I won't have that much dedicated coding time for quite some time. But like I explained earlier, these are simple bugs actually - including some smart pointer related issues and leaks...
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11738
Joined: Tue Mar 25, 2008 8:40 am

Re: Populate AI Traffic with real traffic

Postby Hooray » Sat Nov 24, 2012 3:41 pm

Just a little update that could turn out to be relevant at some point: https://gitorious.org/fg/flightgear/com ... 399bfa83d6
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11738
Joined: Tue Mar 25, 2008 8:40 am

Re: Populate AI Traffic with real traffic

Postby F-JJTH » Sat Nov 24, 2012 6:59 pm

Indeed ! It seems that HLA and AI are now linked :) It's a really good new.
Isn't it the good moment to speak about FGAIS on devel-list ? or at least to Mathias ?
User avatar
F-JJTH
 
Posts: 697
Joined: Fri Sep 09, 2011 11:02 am

Re: Populate AI Traffic with real traffic

Postby Hooray » Sat Nov 24, 2012 7:10 pm

it would definitely seem possible now to re-implement FGAIS on top of HLA - the fgais code is certainly a good demo/test case.
Which also means that we don't necessarily need to look into improving the current implementation further to remove existing bugs.

The HLA route is very promising (superior!) because it would also allow us to transmit derivative information instead of absolute positions, to reduce the traffic overhead.
It is definitely much better than the fgms approach. Even though that was already better than the original pure client-side implementation.
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11738
Joined: Tue Mar 25, 2008 8:40 am

Re: Populate AI Traffic with real traffic

Postby HJ1AN » Thu May 15, 2014 1:50 am

Apologies for bumping an old thread, but interested to know what became of this?
User avatar
HJ1AN
 
Posts: 377
Joined: Sat Jul 25, 2009 4:45 am
Callsign: HJ888
Version: 3.4
OS: OS X

Re: Populate AI Traffic with real traffic

Postby F-JJTH » Thu May 15, 2014 9:08 am

The idea has been abandonned.

The reason: I finally changed my mind, and agreed that FlightGear is a simulator. This means that the FG environment (like AI traffic in our case) must be generating/handled/managed by FlightGear, not by an external source.

Regards,
Clément
User avatar
F-JJTH
 
Posts: 697
Joined: Fri Sep 09, 2011 11:02 am

Re: Populate AI Traffic with real traffic

Postby HJ1AN » Thu May 15, 2014 10:27 am

That's a shame. It'd be great if there is an option to track just one or two flights though, and follow it around or formation flight
User avatar
HJ1AN
 
Posts: 377
Joined: Sat Jul 25, 2009 4:45 am
Callsign: HJ888
Version: 3.4
OS: OS X

Re: Populate AI Traffic with real traffic

Postby Johan G » Thu May 15, 2014 9:40 pm

In addition it made for a huge surge in the bandwidth the servers had to stand up against. :|

EDIT: I might have been a bit confused doing that edit. I guess this topic deals with AI traffic locally, while the one I was thinking of injected real traffic into the MP system.
Last edited by Johan G on Thu May 15, 2014 11:48 pm, edited 1 time in total.
Low-level flying — It's all fun and games till someone looses an engine. (Paraphrased from a YouTube video)
Improving the Dassault Mirage F1 (Wiki, Forum, GitLab. Work in slow progress)
Johan G
Moderator
 
Posts: 5691
Joined: Fri Aug 06, 2010 5:33 pm
Location: Sweden
Callsign: SE-JG
IRC name: Johan_G
Version: 3.0.0
OS: Windows 7, 32 bit

PreviousNext

Return to AI Traffic

Who is online

Users browsing this forum: No registered users and 3 guests