Tyre Physics Model (ISI TGM Tyre Tool) Video, have people seen this?

Discussion in 'General Discussion' started by Spinelli, May 27, 2012.

  1. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    Makes total sense
     
  2. PLAYLIFE

    PLAYLIFE Registered

    Joined:
    Dec 26, 2010
    Messages:
    743
    Likes Received:
    126
    Yep, provided your interpolation and extrapolation algorithms are appropriate in terms of resolution.
     
  3. jubuttib

    jubuttib Registered

    Joined:
    Dec 16, 2011
    Messages:
    934
    Likes Received:
    7
    I'd also like to point out that accuracy is the enemy of dynamic systems in real-time simulation. Lookup tables can achieve much more accurate data while remaining fast enough to use in real-time.

    With lookup tables the trade-off isn't so much accuracy, but range. You can generate as accurate results as you want into the lookup table by making the calculations more accurate, without having to worry about real-time performance. The trade-off is that once calculated you can't get results from outside the bounds of the table. This is only really a problem if the initial boundaries aren't specified to handle the whole range of situations encountered.
     
  4. jubuttib

    jubuttib Registered

    Joined:
    Dec 16, 2011
    Messages:
    934
    Likes Received:
    7
    If both systems run on same boundaries and resolution, and the iteration cycle does go through the range properly, then in theory at least there is no need for interpolation or extrapolation.
     
  5. PLAYLIFE

    PLAYLIFE Registered

    Joined:
    Dec 26, 2010
    Messages:
    743
    Likes Received:
    126
    There'd be no need for interpolation if it remains in range, as you said.

    If it's out-of-range then there's a need for extrapolation. Given the obvious lack of 'real data' out of range, the extrapolation algorithm needs to be sound otherwise funky results occur. Difficult to ascertain a sound algorithm when dealing with extrapolation because you effectively have no baseline instead use 'engineering judgement' as we like to call it.
     
  6. jubuttib

    jubuttib Registered

    Joined:
    Dec 16, 2011
    Messages:
    934
    Likes Received:
    7
    A dynamic model doesn't work any better outside its boundaries though, which was one of the points. =)
     
  7. PLAYLIFE

    PLAYLIFE Registered

    Joined:
    Dec 26, 2010
    Messages:
    743
    Likes Received:
    126
    No definitely not, but it can work a lot worse!
     
  8. 88mphTim

    88mphTim racesimcentral.net

    Joined:
    Sep 23, 2010
    Messages:
    10,840
    Likes Received:
    314
    IMHO that's one of the best ways to define a simulation.

    It's sometimes not about how realistic you can drive it, but how unrealistically you can't drive it.
     
  9. TJones

    TJones Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,074
    Likes Received:
    257
    Good point, this irritated me in the past, when some guys, uses some totaly unrealistic setup adjustments, and therefor take advantage of the weaknesses of the physics engine.
    Not to blame the guys who took the advantage, but the simulation.
     
  10. afborro

    afborro Registered

    Joined:
    Nov 3, 2010
    Messages:
    17
    Likes Received:
    0
    That is a pretty healthy view anyway from another angle. Assume the worst, not the best case :)

    Even in range as it were there can be issues with the approximations used in any model.

    While I was trying to educate myself on the internals of Brush models and derivatives, grinding away through the math and the derivation with the papers I got on it, you quickly realise the levels of approximations and kind of assumptions made.

    For me it is a reminder once again, and puts things into perpective though that in my view people believe in it a bit too much at times, how accurately things can be replicated in driving sims today when it comes to some of these things, Personally I often take it with a pinch of salt, not even considering lack of real data, the model itself and the limited way it can be applied in realtime is still a big factor and limitation IMO.

    It reminds me of the physics joke I used to use.

    Let's assume a spherical horse on a table with an infinite number of legs of negligible cross section.

    Not that bad, but you know where I am going :)

    It is interesting that a lot of people tend to think of Brush beeing the latest and sexy all dancing tire model, but when you read some of the original publications on the subject you quickly realise it too has issues and approximations/assumptions that make me frown at times.
     
  11. jubuttib

    jubuttib Registered

    Joined:
    Dec 16, 2011
    Messages:
    934
    Likes Received:
    7
    Ah, I thought you were coming at it from the opposite direction! I'm with you there. =)
    "I have a solution, but it only works for spherical chickens in a vacuum."
     
  12. PLAYLIFE

    PLAYLIFE Registered

    Joined:
    Dec 26, 2010
    Messages:
    743
    Likes Received:
    126
    Coincidently, I was listening to an interview with Franke Dernie this morning and he mentioned how their current tyre and aero simulations are not fantastic. The interview was regarding driver confidence and he made a tongue in cheek comment that the 'driver's confidence variable' is impossible to model in simulations lol.
     

Share This Page