There's plenty of interest, even from us c++ gimps I teach using MOgre and C#, so a mogre version would be very much appreciated. In fact just today we were trying to figure out how we could get pathing and avoidance going So post away or PM me and I'll test the sucker out. Thankshandcircus wrote:Probably not of interest as its .net only but I just ported OpenSteer to C# and Mogre - I'll try and post it up somewhere soon. All that mix-in stuff was a bit weird - made me feel a bit dizzy.jacmoe wrote:I am still keen on making OgreSteer happen.
This looks fun. that Buckland book is ace! Nice work
little ai demo (steering behaviours)
-
- Ogre Magi
- Posts: 1266
- Joined: Tue Aug 12, 2003 1:53 am
- Location: Melbourne, Australia
- x 1
- Contact:
-
- Kobold
- Posts: 27
- Joined: Sun Feb 11, 2007 11:18 pm
- Location: Sunny Scotland
sorry handcircus i must of missed your post somehow - sounds like you have done some excellent work, the mixin stuff etc was what put me off using opensteer as i dont really understand it, hope you post your ported version up soon!
jacmoe - have signed up to your forum, but the registration confirmation email seems to be taking a long time to come through
jacmoe - have signed up to your forum, but the registration confirmation email seems to be taking a long time to come through
- jacmoe
- OGRE Retired Moderator
- Posts: 20570
- Joined: Thu Jan 22, 2004 10:13 am
- Location: Denmark
- x 179
- Contact:
You should have received the activation email by now.antarctica2007 wrote:jacmoe - have signed up to your forum, but the registration confirmation email seems to be taking a long time to come through
The forum is run from the sourceforge servers, and they don't allow PHP to access the mail client, for security reasons.
I've made a cron job run on my shell account, sending the queued messages on an hourly basis. A work-around which works, but requires some patience. The CEGUI folks are doing the same thing, AFAIK.
Anyway, looking forward to seeing you there.
/* Less noise. More signal. */
Ogitor Scenebuilder - powered by Ogre, presented by Qt, fueled by Passion.
OgreAddons - the Ogre code suppository.
Ogitor Scenebuilder - powered by Ogre, presented by Qt, fueled by Passion.
OgreAddons - the Ogre code suppository.
-
- Kobold
- Posts: 27
- Joined: Sun Feb 11, 2007 11:18 pm
- Location: Sunny Scotland
Hmm been having a rethink - as i mentioned in an earlier post i was going to have the physics&ai tightly coupled as i wanted it to be able to cope with obstacles of just about any shape,so i was going to use OgreNewts convex hull and do ray/tri tests etc, but im now not sure thats the best idea as it forces people to use the same physics implementation as i do.
as spannerman says i guess it would be better to decouple the ai layer from the movement/physics layer, which would make it more portable; ideally it would be so that the user just added the .h/.cpp or the .dll to their project and didnt really have to worry too much about how it works.
also from what spannerman said i think i should not bother trying to make a contrived wall avoidance behaviour work, and instead improve the obstacle avoidance to be more generic & dynamic,
although the basics could be achieved with simple RaySceneQueries vs bounding boxes/spheres, perhaps i would be better off with something like coldet or OgreOpcode (i havent really checked these out much yet though) as a halfway house solution, removing the depenancy on any one particular physics library, but allowing them to be used at the users disgression- of course this adds a new dependancy but hopefully a less obtrusive one to the user
still no email - i checked i hadnt used the wrong on by trying to sign up again and it says its definitely taken though, i dont suppose you can process a confimation email by hand if its still not through soon?
as spannerman says i guess it would be better to decouple the ai layer from the movement/physics layer, which would make it more portable; ideally it would be so that the user just added the .h/.cpp or the .dll to their project and didnt really have to worry too much about how it works.
also from what spannerman said i think i should not bother trying to make a contrived wall avoidance behaviour work, and instead improve the obstacle avoidance to be more generic & dynamic,
although the basics could be achieved with simple RaySceneQueries vs bounding boxes/spheres, perhaps i would be better off with something like coldet or OgreOpcode (i havent really checked these out much yet though) as a halfway house solution, removing the depenancy on any one particular physics library, but allowing them to be used at the users disgression- of course this adds a new dependancy but hopefully a less obtrusive one to the user
still no email - i checked i hadnt used the wrong on by trying to sign up again and it says its definitely taken though, i dont suppose you can process a confimation email by hand if its still not through soon?
- jacmoe
- OGRE Retired Moderator
- Posts: 20570
- Joined: Thu Jan 22, 2004 10:13 am
- Location: Denmark
- x 179
- Contact:
I will check your account, and activate you manually if need be.
About OgreSteer:
I think it should be a steering/pathing library, not relying on any particular library for collision detection, etc.
I had in mind something like a PathAffector (just like particle systems, actually).
If a collision happens, the path is affected. And so on.
Some ready-made affectors could be provided. One is those would of course be a plain old raycast.
Others would be hooks into OgreNewt, NxOgre, OgreBullet, etc.
IMO, the library should be using Ogre - that's why it's a wrapper.
Wrapping not only OpenSteer, but also MicroPather.
And our very own home-made code..
Some helper code/classes will probably be made, like state machines and so on, but I don't think it should be an AI library.
Well, I guess I should move this discussion to the OgreSteer forum ..
About OgreSteer:
I think it should be a steering/pathing library, not relying on any particular library for collision detection, etc.
I had in mind something like a PathAffector (just like particle systems, actually).
If a collision happens, the path is affected. And so on.
Some ready-made affectors could be provided. One is those would of course be a plain old raycast.
Others would be hooks into OgreNewt, NxOgre, OgreBullet, etc.
IMO, the library should be using Ogre - that's why it's a wrapper.
Wrapping not only OpenSteer, but also MicroPather.
And our very own home-made code..
Some helper code/classes will probably be made, like state machines and so on, but I don't think it should be an AI library.
Well, I guess I should move this discussion to the OgreSteer forum ..
/* Less noise. More signal. */
Ogitor Scenebuilder - powered by Ogre, presented by Qt, fueled by Passion.
OgreAddons - the Ogre code suppository.
Ogitor Scenebuilder - powered by Ogre, presented by Qt, fueled by Passion.
OgreAddons - the Ogre code suppository.
-
- Kobold
- Posts: 27
- Joined: Sun Feb 11, 2007 11:18 pm
- Location: Sunny Scotland
It looks like im not going to be able to work on this much atm so i thought i would release the source "as is" in case its of use to anybody,
http://homepages.nildram.co.uk/~ant2k3/projects.html
even though it doesnt actually use any ogrenewt physics body's atm its still requires it for compliation etc- this needs to be sorted out in project settings plus modifying the includes etc
http://homepages.nildram.co.uk/~ant2k3/projects.html
even though it doesnt actually use any ogrenewt physics body's atm its still requires it for compliation etc- this needs to be sorted out in project settings plus modifying the includes etc
-
- Halfling
- Posts: 43
- Joined: Wed Apr 05, 2006 8:00 pm
- x 2
Cool, it basically there but really wanna tidy it up a little - working on quite a few projects at the mo and have no time, but as there is interest i'll prioritiseVectrex wrote:There's plenty of interest, even from us c++ gimps I teach using MOgre and C#, so a mogre version would be very much appreciated. In fact just today we were trying to figure out how we could get pathing and avoidance going So post away or PM me and I'll test the sucker out. Thankshandcircus wrote:Probably not of interest as its .net only but I just ported OpenSteer to C# and Mogre - I'll try and post it up somewhere soon. All that mix-in stuff was a bit weird - made me feel a bit dizzy.jacmoe wrote:I am still keen on making OgreSteer happen.
This looks fun. that Buckland book is ace! Nice work
Ive not ported the demo framework, just opensteer itself, so i'll plonk in a few demos and try to get that up in the next week. Might make it eihort dependent when it comes out too (which should be this weekend, according to Bekas).
HandCircus Game Studio - http://www.handcircus.com - Creators of Rolando and Rolando 2
- SpannerMan
- Gold Sponsor
- Posts: 446
- Joined: Fri May 02, 2003 10:05 am
- Location: UK
- Contact:
Re: little ai demo (steering behaviours)
Hey, did anybody manage to retain a copy of this sourcode for antarctica2007's Steering demo? All the links are dead.
- Danoli3
- Halfling
- Posts: 53
- Joined: Wed Jul 29, 2009 8:25 am
- Location: Sydney, Australia
- x 1
- Contact:
Re: little ai demo (steering behaviours)
Also interested if anyone has source?
Studying BSc in Games Development @ the University of Technology of Sydney Australia
Ogre Projects so far: Mall Rush
Finished a Diploma in Games Development 2009
Ogre Projects so far: Mall Rush
Finished a Diploma in Games Development 2009
-
- Kobold
- Posts: 29
- Joined: Mon Apr 02, 2007 4:42 pm
Re: little ai demo (steering behaviours)
steersuite http://www.magix.ucla.edu/steersuite/ is another steering behaviors project which provoids more test enviroments.
- Danoli3
- Halfling
- Posts: 53
- Joined: Wed Jul 29, 2009 8:25 am
- Location: Sydney, Australia
- x 1
- Contact:
Re: little ai demo (steering behaviours)
awesome thanks yaoyansi
Studying BSc in Games Development @ the University of Technology of Sydney Australia
Ogre Projects so far: Mall Rush
Finished a Diploma in Games Development 2009
Ogre Projects so far: Mall Rush
Finished a Diploma in Games Development 2009