Chris
Self taught here. Just read, think, and experiment a lot. The physics work
I'm doing is going into a free, online radio controlled car game that'll be
released someday :-) Will let you all know about it when it happens.
Most of the development work I do for the project is on simulated full size
cars, so it works with about any car you want to model. This is done with an
OpenGL 3-D graphics system that I might try to chase Ruud Van Gaal's Racer down
with someday ;-) Although he's far, far more into 3-D graphics than I am right
now!
I'm not sure how my level of vehicle dynamics knowledge compares to others
around these parts, but I seem to hold my own in some of the discussions ok.
I'm getting to the point where suspension kinematics are included (using ball
bearing locations, arm lengths, etc..) not something too many other folks (in
the video game world, anyway) have done yet.
Tire modelling isn't something I've gotten into much yet, although I've
written a little brush model that might one day allow reasonably accurate force
curves to be developed based on some simple tire physical characteristics for
less than $30,000 ;-) . Not heavy duty engineering stuff, but pretty cool
anyway. Pacejka and Raadt models are not my strong point yet though, haven't
read or played around enough with them yet. Too busy with the other parts.
If you or anybody else want a screenshot or two (not very impressive, be
warned), email me and I'll send you something.
Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com
Mine's for fun, a small attempt at how far I can get this to go, a
more sim-like feel like GPL with an really open approach, so I don't
have to do all the car and track work. :) And besides, such an
approach makes the product live longer, I think. It needs to, because
it takes a lot of work, I've noticed!
There are screenshots and all that on my site, of which 2 include a
new Subaru, created by one of the Racer forum regular's, Mitch. Really
done for Colin McRae Rally 2, but Racer had the premiere. :)
Not really *** knowledge, I'm peddling along, buying books along
the way, and I learned most I think from the talk here at RAS. Still,
the books give pretty explaining pictures. ;)
My real work is programming applications for TV-shows, quizzes, that
sort of stuff. So the accent is a bit on graphics sometimes, with the
notion that the physics is still very much required to get a cultlike
crowd to avoid drowning in all the terrific carsim titles that
surround us.
Ruud van Gaal, GPL Rank +53.25
Pencil art : http://www.racesimcentral.net/
Free car sim : http://www.racesimcentral.net/
Based on the recent level of posting, it looks to me like vehicle dynamics
will soon take over ras! Time for everyone to quit playing around and
crack the text books<grin>.
-- Doug Milliken
Milliken Research Associates Inc.
> Chris
| Never happen Doug....it's been over 10 years since I've read a
book,....a
| side effect of Narcolepsy...open cover see page..zzzzzzzzzz.zzzzzzzzz
| zzzzzzzzzzz
| dave henrie
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.265 / Virus Database: 137 - Release Date: 7/18/2001
Any suggestions for which text book would be relevant in a case like this?
<grin>
Ash :)
|
| Based on the recent level of posting, it looks to me like vehicle dynamics
| will soon take over ras! Time for everyone to quit playing around and
| crack the text books<grin>.
|
| -- Doug Milliken
| Milliken Research Associates Inc.
|
|
I blame Ruud van Gaal! Personally I've tried writing a vehicle simulator in
the past but got stumped trying to find information. This time I happened
across Ruud's Racer site, which mentioned this newsgroup. So although I'm
new to the newsgroup, I'm not new to attempting vehicle simulators. Maybe
it's a similar story for some of the others?
Mainly as a learning aid. I've worked in the games industry for a number of
years, but have never worked on a driving game. I'm hoping that writing a
decent simulation will help push me into a role where I can do this.
Besides, it's a lot of fun :)
Nothing much to look at yet. The driving model is starting to come together
pretty well, but the graphics are just functional for the moment. The source
code resembles spaghetti (sp?) since this is my first attempt at this. Once
I'm convinced I have the model right, I'll be rewriting the whole thing to
make it tidier, more extendible etc. At this point I'll probably post
binaries and source on my homepage for download.
Self taught car dynamics from info I've found on the web, this newsgroup
etc. Although I do have a strong maths/physics background which helps a
little. The good thing about vehicle dynamics is that you only need a small
amount of information to get a simple simulator working. But if you want you
can go more in depth and work towards a full simulation, which can get
pretty complex. So you can work in stages - start with something simple and
build on it. Every new thing you add gives a little more realism to the
simulation
> Self taught car dynamics from info I've found on the web, this newsgroup
> etc. Although I do have a strong maths/physics background which helps a
> little. The good thing about vehicle dynamics is that you only need a small
> amount of information to get a simple simulator working. But if you want you
> can go more in depth and work towards a full simulation, which can get
> pretty complex. So you can work in stages - start with something simple and
> build on it. Every new thing you add gives a little more realism to the
> simulation
I'm not in English academic language mode today, but I hope you get
what I mean, and the final question is then:
Have you thought about this when making car simulations more and more
complex ?
--
Olav K. Malmin
remove .spam when replying
that's a very good point, and indeed it's true. The stuff in the vehicle
models can be sonewhat grouped in terms of influence. The basic core (a
rigid body with a few contact points) is the same, and then you may
proceed to lower order parameters that have a smaller and smaller effect
on the overall behaviour. There definitely exists a point of overkill,
where introducing additional features doesn't bring anything new, but
can indeed start throwing off the calculations slightly because of the
additional uncertainty.
I don't think we're quite that far yet, since anything new we're adding
at this point has a significant effect on car handling and is as such
significant, but once the features are being added just because they can
be while not bringing anything new to the experience, then indeed one
needs to stop.
And there's more than just the reason of uncertainty for doing so :).
-Gregor
> I just think I would throw in a little point in that respect. In every
> course I had at the university about predictions and modelling we have
> learned that the greater number of parameters into a model also brings
> more inaccuracy to the model. So less parametres gives a very
> simplified result but with little uncertainty (don't really know what
> this is called in english) and more parametres gives a result that
> might be more corrrect, but with a higher degree of uncertainty.
> I'm not in English academic language mode today, but I hope you get
> what I mean, and the final question is then:
> Have you thought about this when making car simulations more and more
> complex ?
> --
> Olav K. Malmin
> remove .spam when replying
>I just think I would throw in a little point in that respect. In every
>course I had at the university about predictions and modelling we have
>learned that the greater number of parameters into a model also brings
>more inaccuracy to the model. So less parametres gives a very
>simplified result but with little uncertainty (don't really know what
>this is called in english) and more parametres gives a result that
>might be more corrrect, but with a higher degree of uncertainty.
However, I'm noticing that the things I'm adding at this moment are
secondary effects; detailed versions of things that were mostly in
already (like adding caster camber to the steering code). The effects
are more difficult to feel, but some people notice the subtle points.
So I don't think it really accounts in this case, not entirely at
least. But ofcourse, there's a limit; you can simulate all you want,
but as all of it is just a model there's no use doing every nitty
gritty thing, since the model itself is intrinsically inaccurate,
because it is just modeling a highly complex parallel vehicle.
Ruud van Gaal, GPL Rank +53.25
Pencil art : http://www.marketgraph.nl/gallery/
Free car sim : http://www.marketgraph.nl/gallery/racer/
Of course you've got to know when to stop :)
When developing a game, you have a limited amount of time available, so at
some point you have to say the model is good enough and go with it. You will
also have people working around you on other parts of the game - levels, AI,
artwork etc. They will need a physics model right near the beginning to
begin their work - they can't wait until you say it's finished. This is the
main reason for the approach I've mentioned. You give them a simple model at
the beginning, and then release the modifications to them as they become
available.
At some point the parameters that you're adding in will begin to have little
or no noticable effect on your model, and then somebody has to make an
educated decision and say that's it.
Of course it also depends on the type of game you are developing - an arcade
racer will not need as many parameters as a sim. You may choose to ignore
camber angle, or Ackerman angle, or even only have one gear.
OK, here's the controversial bit:
As for uncertainty, the accuracy of the model may be theoretically
uncertain, but providing it feels accurate, that's sufficient for a game.
I'm not writing a simulation for safety testing - nobody's life is at risk
here ;-)
Mike
<snip uncertainty theory>
Games ?
I thought we discussed *** theory here :)
--
Olav K. Malmin
remove .spam when replying
