Re: [Flightgear-devel] Question on YASim ....
Jon S Berndt wrote: On Thu, 2 Jan 2003 14:54:15 -0800 (PST) Gene Buckle [EMAIL PROTECTED] wrote: My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I don't know of any other simulator that could do this, but I don't know if it's practical either. I think it would be fantastic if it was possible. Hmmm. Cool idea. Of course, it would be sort of hard to modify table lookups for a complex coefficient, but if you start with a simple model you could play with coefficients in real time I think by using a property browser. Hmm. Hmm. Hmm. It might come in handy for CG and reff. pt. locations also ... Erik ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
RE: [Flightgear-devel] Question on YASim ....
I like this. I am using evolutionary computing methods to modify flight behaviors, but they could also be applied to optimize aircraft models for whatever property that you want to optimize. The method is based on biological evolution (survival of the fittest and all that). This is how birds learned to fly and tuned themselves to their environment (why albatrosses have a different aerodynamic form and set of behaviors than buzzards, for example). With JSBSim you can parallelize and run simulations off-line on a supercomputer (no visualization, just the FDM), and then extract the optimized values. Anybody want to help me with this? Mark Boslough Evolutionary Computing Department 9216 PO Box 5800, MS 0318 Sandia National Laboratories Albuquerque, NM 87185-0318 -Original Message- From: Tony Peden [mailto:[EMAIL PROTECTED]] Sent: Thursday, January 02, 2003 5:26 PM To: FGFS Devel Subject: Re: [Flightgear-devel] Question on YASim On Thu, 2003-01-02 at 14:58, Jon S Berndt wrote: On Thu, 2 Jan 2003 14:54:15 -0800 (PST) Gene Buckle [EMAIL PROTECTED] wrote: My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I don't know of any other simulator that could do this, but I don't know if it's practical either. I think it would be fantastic if it was possible. Hmmm. Cool idea. Of course, it would be sort of hard to modify table lookups for a complex coefficient, but if you start with a simple model you could play with coefficients in real time I think by using a property browser. Hmm. Hmm. Hmm. Tony? They are modifiable now to a certain extent. Each coefficient has a gain and bias term that is applied like: actual coefficient = gain*( file coefficient ) + bias If the coefficient is given as a table in the file, then the file coefficient value above is the result of the table lookup. These can be modified on the fly via the property tree, go to: /fdm/jsbsim/aero/buildup/axis/coeff name This is a very effective method for tweaking one or two coeff's. There is currently no way to dump the gain and bias values or rewrite the config file though so you wouldn't want tweak too many at once. Jon Coordinator [EMAIL PROTECTED] JSBSim Project www.JSBSim.org ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel -- Tony Peden [EMAIL PROTECTED] ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
RE: [Flightgear-devel] Question on YASim ....
Boslough, Mark B writes: I like this. I am using evolutionary computing methods to modify flight behaviors, but they could also be applied to optimize aircraft models for whatever property that you want to optimize. The method is based on biological evolution (survival of the fittest and all that). This is how birds learned to fly and tuned themselves to their environment (why albatrosses have a different aerodynamic form and set of behaviors than buzzards, for example). With JSBSim you can parallelize and run simulations off-line on a supercomputer (no visualization, just the FDM), and then extract the optimized values. Anybody want to help me with this? Not that I have extra time for anything new, but here's what I did for my grad school project (and yes, I paid my $1 so I am authorized to use the cool it works with Linux logo, but I'm probably dating myself there, although that may not be legal yet in all 50 states.) http://www.menet.umn.edu/~curt/research/ We didn't call it an evolutionary method though, we just made a WAG and tested to see if we were any better off than before. :-) Essentially our *very large* search space was partitioned statistically which means if every process is off making WAG's, their's a good chance they each are exploring different portions of the overall search space. This sounds pretty hokey at first, but actually works surprisingly well. Each process was running a self contained AI search algorithm and was oblivious to what all the other processes were doing. The first process to find a solution won. The assumption here is that the most quickly found solution is probably the shortest and most optimal. For other applications you might need a different metric ... i.e. the process that hits the ground last wins ... I was ganging up 25 linux machines to find solutions. Earlier versions were developed using a honest to goodness, real life super computer, but it was a pain in the butt to get time on it, and everything was batched and ran lower priority than the paying customers, so it was usually faster to run on your own PC(s). A nice feature of what I was doing was that none of the processes needed to communicate with each other other than to call BINGO!. I'd be happy to answer more specific questions about what I did, and you're welcome to browse the code and help yourself to any of it. Unless otherwise stated, assume GPL ... Curt. -- Curtis Olson IVLab / HumanFIRST Program FlightGear Project Twin Cities[EMAIL PROTECTED] [EMAIL PROTECTED] Minnesota http://www.menet.umn.edu/~curt http://www.flightgear.org ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
On Fri, 3 Jan 2003 13:45:31 -0700 Boslough, Mark B [EMAIL PROTECTED] wrote: This is how birds learned to fly and tuned themselves to their environment (why albatrosses have a different aerodynamic form and set of behaviors than buzzards, for example). With JSBSim you can parallelize and run simulations off-line on a supercomputer (no visualization, just the FDM), and then extract the optimized values. Anybody want to help me with this? !! Sounds like this has a pretty high Drool Factor. Alas, my wife says it will be at least next year until she will let me buy a supercomputer, after diaper expenses start to subside. Seriously, what kind of help are you looking for? I am actually working on some additional JSBSim features which have gone unfinished for too long, so I don't think I have much time, but I am curious. Are you doing formal or informal work? Jon ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Jon S Berndt writes: Sounds like this has a pretty high Drool Factor. Alas, my wife says it will be at least next year until she will let me buy a supercomputer, after diaper expenses start to subside. Who needs a supercomputer? Buy 15 cheap $200 boxes and put them in a cluster. All the best, David -- David Megginson, [EMAIL PROTECTED], http://www.megginson.com/ ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
RE: [Flightgear-devel] Question on YASim ....
And heat your house too! ;-) -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of David Megginson Sent: January 3, 2003 4:49 PM To: [EMAIL PROTECTED] Subject: Re: [Flightgear-devel] Question on YASim Jon S Berndt writes: Sounds like this has a pretty high Drool Factor. Alas, my wife says it will be at least next year until she will let me buy a supercomputer, after diaper expenses start to subside. Who needs a supercomputer? Buy 15 cheap $200 boxes and put them in a cluster. All the best, David -- David Megginson, [EMAIL PROTECTED], http://www.megginson.com/ ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Jon S Berndt writes: On Fri, 3 Jan 2003 17:49:00 -0500 David Megginson [EMAIL PROTECTED] wrote: Jon S Berndt writes: Sounds like this has a pretty high Drool Factor. Alas, my wife says it will be at least next year until she will let me buy a supercomputer, after diaper expenses start to subside. Who needs a supercomputer? Buy 15 cheap $200 boxes and put them in a cluster. That's about the diaper/pull-ups budget for the coming year. No joke! You could probably heat your house with these too. Curt. -- Curtis Olson IVLab / HumanFIRST Program FlightGear Project Twin Cities[EMAIL PROTECTED] [EMAIL PROTECTED] Minnesota http://www.menet.umn.edu/~curt http://www.flightgear.org ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
RE: [Flightgear-devel] Question on YASim ....
A good example of parallelism. -Original Message- From: Michael Bonar [mailto:[EMAIL PROTECTED]] Sent: Friday, January 03, 2003 4:11 PM To: [EMAIL PROTECTED] Subject: RE: [Flightgear-devel] Question on YASim And heat your house too! ;-) -Original Message- From: Curtis L. Olson [mailto:[EMAIL PROTECTED]] Sent: Friday, January 03, 2003 4:15 PM To: [EMAIL PROTECTED] Subject: Re: [Flightgear-devel] Question on YASim You could probably heat your house with these too. Curt. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
On Fri, 3 Jan 2003 16:20:36 -0700 Boslough, Mark B [EMAIL PROTECTED] wrote: A good example of parallelism. From: Michael Bonar [mailto:[EMAIL PROTECTED]] And heat your house too! ;-) From: Curtis L. Olson [mailto:[EMAIL PROTECTED]] You could probably heat your house with these too. Hmmm. That might explain why the babies room is always so warm ... :-P Mark: I'd still be interested in hearing more about your project to whatever degree you can discuss. Jon ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
RE: [Flightgear-devel] Question on YASim ....
That's pretty much what I'm using, but more than 15 processors: http://www.cs.sandia.gov/cplant/ Mark Who needs a supercomputer? Buy 15 cheap $200 boxes and put them in a cluster. All the best, David ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
RE: [Flightgear-devel] Question on YASim ....
Communication helps, just as in real evolution. You want the survivors to be able to have sex with each other (crossover of genes). But the communication/computation ratio is usually very low for what I am doing. Not that I have extra time for anything new, but here's what I did for my grad school project (and yes, I paid my $1 so I am authorized to use the cool it works with Linux logo, but I'm probably dating myself there, although that may not be legal yet in all 50 states.) http://www.menet.umn.edu/~curt/research/ We didn't call it an evolutionary method though, we just made a WAG and tested to see if we were any better off than before. :-) Essentially our *very large* search space was partitioned statistically which means if every process is off making WAG's, their's a good chance they each are exploring different portions of the overall search space. This sounds pretty hokey at first, but actually works surprisingly well. Each process was running a self contained AI search algorithm and was oblivious to what all the other processes were doing. The first process to find a solution won. The assumption here is that the most quickly found solution is probably the shortest and most optimal. For other applications you might need a different metric ... i.e. the process that hits the ground last wins ... I was ganging up 25 linux machines to find solutions. Earlier versions were developed using a honest to goodness, real life super computer, but it was a pain in the butt to get time on it, and everything was batched and ran lower priority than the paying customers, so it was usually faster to run on your own PC(s). A nice feature of what I was doing was that none of the processes needed to communicate with each other other than to call BINGO!. I'd be happy to answer more specific questions about what I did, and you're welcome to browse the code and help yourself to any of it. Unless otherwise stated, assume GPL ... Curt. -- Curtis Olson IVLab / HumanFIRST Program FlightGear Project Twin Cities[EMAIL PROTECTED] [EMAIL PROTECTED] Minnesota http://www.menet.umn.edu/~curt http://www.flightgear.org ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
RE: [Flightgear-devel] Question on YASim ....
And that's not even counting the ones for your kids. All the best, David note to self: make sure webcam is disconnected. someone knows more than they should. :-P smime.p7s Description: application/pkcs7-signature
[Flightgear-devel] Question on YASim ....
Andy, is it technically possible to fiddle with the model parameters in real-time? g. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Gene Buckle wrote: Andy, is it technically possible to fiddle with the model parameters in real-time? Not easily. Changing the parameters requires a re-solution, which can take a second or two for aircraft with lots of elements like the 747. So it would have to be done a little bit at a time over many frames, and probably involve a throttler gadget to keep the frame rate high enough. On an SMP system, you could just spawn a thread to do it and it would work great, of course. But on a uniprocessor, even threading would take half the CPU and performance would drop by 50% for a few seconds while the solution popped out. Is there something in particular you want to control at runtime? Support could probably be added per-device. Andy -- Andrew J. RossNextBus Information Systems Senior Software Engineer Emeryville, CA [EMAIL PROTECTED] http://www.nextbus.com Men go crazy in conflagrations. They only get better one by one. - Sting (misquoted) ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Gene Buckle wrote: Andy, is it technically possible to fiddle with the model parameters in real-time? Not easily. Changing the parameters requires a re-solution, which can take a second or two for aircraft with lots of elements like the 747. So it would have to be done a little bit at a time over many frames, and probably involve a throttler gadget to keep the frame rate high enough. On an SMP system, you could just spawn a thread to do it and it would work great, of course. But on a uniprocessor, even threading would take half the CPU and performance would drop by 50% for a few seconds while the solution popped out. Is there something in particular you want to control at runtime? Support could probably be added per-device. (thanks for the quick response) My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I imagine it would save a _lot_ of time if the edit parameter file, run fgfs, test, re-edit cycle could be reduced to run fgfs, tweak in-flight, dump new parameter file. I've got _way_ too many irons in the fire to try to tackle this myself, but if YASim could listen on a TCP port for model wrench commands, an external tool could be written (gui or text) that would allow a user to tweak the currently running model in-flight and then have the tool dump the changed version of the parameter file to disk after the session was complete. YASim itself wouldn't have to do much work other than update the requested parameter and re-solve(?) the model. The 1 or 2 second pause wouldn't be any big deal (I wouldn't think) because it would be an expected part of the design phase. If you wanted to add some kind of temporary solver, you could do a test solve and then send an error code to the model wrench and go back to the original (or prior) if the solver dies. I don't know of any other simulator that could do this, but I don't know if it's practical either. I think it would be fantastic if it was possible. Is this too far afield? g. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Off the top of my head, I believe selecting reinit from the file menu causes the instance of the FDM class to be deleted and re-created. This means that it reloads the parameter file and recomputes a solution. That seems like a decent middle of the road solution. You make your fdm changes in a separate window/editor and then do reset to reload them. You have to start again from the ground, but at least you skip the overhead of having to reload flightgear from scratch. This appears to work right now. Regards, Curt. Gene Buckle writes: (thanks for the quick response) My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I imagine it would save a _lot_ of time if the edit parameter file, run fgfs, test, re-edit cycle could be reduced to run fgfs, tweak in-flight, dump new parameter file. I've got _way_ too many irons in the fire to try to tackle this myself, but if YASim could listen on a TCP port for model wrench commands, an external tool could be written (gui or text) that would allow a user to tweak the currently running model in-flight and then have the tool dump the changed version of the parameter file to disk after the session was complete. YASim itself wouldn't have to do much work other than update the requested parameter and re-solve(?) the model. The 1 or 2 second pause wouldn't be any big deal (I wouldn't think) because it would be an expected part of the design phase. If you wanted to add some kind of temporary solver, you could do a test solve and then send an error code to the model wrench and go back to the original (or prior) if the solver dies. I don't know of any other simulator that could do this, but I don't know if it's practical either. I think it would be fantastic if it was possible. Is this too far afield? g. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel -- Curtis Olson IVLab / HumanFIRST Program FlightGear Project Twin Cities[EMAIL PROTECTED] [EMAIL PROTECTED] Minnesota http://www.menet.umn.edu/~curt http://www.flightgear.org ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
On Thu, 2 Jan 2003 14:54:15 -0800 (PST) Gene Buckle [EMAIL PROTECTED] wrote: My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I don't know of any other simulator that could do this, but I don't know if it's practical either. I think it would be fantastic if it was possible. Hmmm. Cool idea. Of course, it would be sort of hard to modify table lookups for a complex coefficient, but if you start with a simple model you could play with coefficients in real time I think by using a property browser. Hmm. Hmm. Hmm. Tony? Jon Coordinator [EMAIL PROTECTED] JSBSim Project www.JSBSim.org ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
On Thu, 2 Jan 2003, Jon S Berndt wrote: On Thu, 2 Jan 2003 14:54:15 -0800 (PST) Gene Buckle [EMAIL PROTECTED] wrote: My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I don't know of any other simulator that could do this, but I don't know if it's practical either. I think it would be fantastic if it was possible. Hmmm. Cool idea. Of course, it would be sort of hard to modify table lookups for a complex coefficient, but if you start with a simple model you could play with coefficients in real time I think by using a property browser. Hmm. Hmm. Hmm. If the table is in the parameter file, there shouldn't be any reason why a designer couldn't change it. Whether or not they know what they're doing of course is an exercise left to the reader. :) With a flexible client, anything should be possible if the FDM can handle it. g. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Jon S Berndt writes: Of course, it would be sort of hard to modify table lookups for a complex coefficient, Not to hard to do if you think of the table as being samples along a curve and have a gui manipulated spline based curve editor. Norman ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Gene Buckle wrote: My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I imagine it would save a _lot_ of time if the edit parameter file, run fgfs, test, re-edit cycle could be reduced to run fgfs, tweak in-flight, dump new parameter file. Ah, got it. Actually, that feature could be expressed as simply as make YASim load its configuration out of the property tree like everything else does. Then you'd just change the properties using whatever interface you want and select reset from the menu. This is definitely the way it should work. I've been slow to do it only out of laziness. Re-writing the existing aircraft definitions in property XML would be a pain. You're right, the N second pause for the solver would be just fine for this application. I was thinking you wanted to do something like variable-camber wings or whatnot and needed realtime control over things that are currently constants to the solver. Although it's worth pointing out that the command line yasim program goes a long way toward reducing the tedium involved with getting an aircraft in the air. Most of the big configuration bugs can be found and fixed before you ever run fgfs, although admittedly interpreting the solution report takes a little practice. Andy -- Andrew J. RossNextBus Information Systems Senior Software Engineer Emeryville, CA [EMAIL PROTECTED] http://www.nextbus.com Men go crazy in conflagrations. They only get better one by one. - Sting (misquoted) ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
On Thu, 2 Jan 2003, Norman Vine wrote: Jon S Berndt writes: Of course, it would be sort of hard to modify table lookups for a complex coefficient, Not to hard to do if you think of the table as being samples along a curve and have a gui manipulated spline based curve editor. Now THAT would be cool. g. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
Although it's worth pointing out that the command line yasim program goes a long way toward reducing the tedium involved with getting an aircraft in the air. Most of the big configuration bugs can be found and fixed before you ever run fgfs, although admittedly interpreting the solution report takes a little practice. I just thought it would be easier for folks to have some kind of graphic tool to develop with. Kind of a FlightGear equivalent to X-Plane's Plane Maker. g. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Question on YASim ....
On Thu, 2003-01-02 at 14:58, Jon S Berndt wrote: On Thu, 2 Jan 2003 14:54:15 -0800 (PST) Gene Buckle [EMAIL PROTECTED] wrote: My thought is that a real-time model wrench might make it easier for people to develop or improve aircraft models. I don't know of any other simulator that could do this, but I don't know if it's practical either. I think it would be fantastic if it was possible. Hmmm. Cool idea. Of course, it would be sort of hard to modify table lookups for a complex coefficient, but if you start with a simple model you could play with coefficients in real time I think by using a property browser. Hmm. Hmm. Hmm. Tony? They are modifiable now to a certain extent. Each coefficient has a gain and bias term that is applied like: actual coefficient = gain*( file coefficient ) + bias If the coefficient is given as a table in the file, then the file coefficient value above is the result of the table lookup. These can be modified on the fly via the property tree, go to: /fdm/jsbsim/aero/buildup/axis/coeff name This is a very effective method for tweaking one or two coeff's. There is currently no way to dump the gain and bias values or rewrite the config file though so you wouldn't want tweak too many at once. Jon Coordinator [EMAIL PROTECTED] JSBSim Project www.JSBSim.org ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel -- Tony Peden [EMAIL PROTECTED] ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel