Difference Tire AI

Discussion in 'Car Modding' started by Rock, Sep 28, 2021.

  1. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Hello,
    My problem is the difference in AI speed when I use different tires ex Medium, Hard.
    I made sure that the .TBC was identical for the two specificities (Hard / Medium) and I did a copy / paste for both and I compared the AI on 6 laps.
    When I put the tires hard for me the AI realizes 1.26.00 but if I put the Medium tires the AI realizes 1.27.00. And he would have preferred the opposite ;-)
    I tried to change my .TGM by the original ones from ISI close to mine but the problem remains the same.
    Do you have any idea on this problem?
     
  2. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    Hi
    As first thing, when You speak about some problem with a mod....You should say which is the mod.
    I mean: which is the car with which You are finding different speed between Human and AI cars?

    First: be sure AI drivers are using the same tire dimensions as Human driver ( Humans are using TGM files, while AI drivers are using TBC files ).
    For this, You can use the Great Online RF2 Tools by Redapg.

    Second: consider that You can edit the AI speed in TBC file changing the value of Rolling resistance ( for Front and Rear tires ).
    So, if You have in your car ( for example ) Hard tires ( TGM for You ), and the max speed of car is.....xx Km/h ( You should make some tests in specific tracks allowing long straight...like Lemans, for example...and without DRS, if You have a F1 car ), You can let AI driving in automatic way in that track with the same tire compound ( Hard in this case...), and without DRS ( temporary deleting it in HDV ).
    Comparing in this way your max speed with that of AI driver, if AI is quicker, try to rise the Rolling Resistance in TBC, or if AI is slower, try to decrease this value, until You get the exact same speed with Human. Make this work also for other compounds.
    All this explanation, of course, is addressed to adjust the max speed...if this is your problem.
    If the problem is not the max speed, but the laptime....well, here You should work around the grip value of tires. For AI drivers, in TBC file, values are Long and Lat grips, while for Human, in TGM, grip values are in the Realtime section, expressed by Static grip, and sliding grip.
     
    Rock likes this.
  3. Coen

    Coen Registered

    Joined:
    Aug 3, 2021
    Messages:
    2
    Likes Received:
    0
    Hi,

    Lately I tried to edit some AI and Human car behaviour as well by edditing the .TBC file. I noticed that in developer mode it is possible to switch between the .TBC and .TGM files for the human operated car as well. However, when I switch to the old (.TBC) tire model, I have 0 grip, while the AI seems to be able to drive normally. Is there another parameter where a difference in grip coefficients for AI and Human is defined, affecting the .TBC file?
     
  4. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    It depends always from the TBC tire grip: for AI drivers, also a poor grip tire may be enough, since it's "the PC" driving the car.
    Normally, if the TBC tire grip si well proportionated to the TGM ( human ) tire grip, You should be able to drive the car on TBC, although with many limits and less correct behavior of the car.
    Therefore, if in your mod dev, when You drive on TBC ( old tire mode ) without grip, probably TBC tires need to be edited, increasing the grip. Unfortunatelly, there is no rule of exact proportion betweem TGM and TBC tire grip.
    Each car needs a specific, balanced set between TGM and TBC.
     
    Coen likes this.
  5. Coen

    Coen Registered

    Joined:
    Aug 3, 2021
    Messages:
    2
    Likes Received:
    0
    Thanks for your answer! More extreme values for Drylonglat seem to work indeed although it does make it very tricky to get realistic behaviour.
     
  6. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    Yes, you are right. But consider that driving a car on TBC tires, in RF2, is no more like in RF1. You have the feeling that the car is driveable….and only on track limits ( TBC tires for human on pitlane or kerbs, or grass, are almost without grip ). Commonly, the most cars in RF2 have the Lat grip in TBC that is from 80 to 85% of Sliding grip in TGM.
    Increasing Lat grip in TBC means to increase proportionally Long grip, and, although their value actually is different from RF1, You should increase in TBC wet grip values proportionally to dry grip values.
     
    Coen likes this.
  7. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Hello Nicolas, Thank you for these excellent explanations. My computer is broken. I'll come back to give you my feedback as soon as it works.
     
  8. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    One question before testing. In TBC file changing the value of "Rolling resistance" works but you should not be careful to force the number to see the difference. But what is the equivalent parameter in the TGM.
     
  9. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    Hello Rock
    To replay to your last question:
    Rolling Resistance in TGM is under [Realtime] section:

    BaseDamperPerUnitArea=(....,....,.....) // values related to the inertia on 3 spatial coordinates

    Increasing values, tyre will be slower; lowering, faster.

    Just to complete the previous discussion about proportions between TGM and TBC:

    TGM ( any compound ) : Static grip / Sliding grip

    for modern F1 is 0.70

    Front / rear ratio 0.9856


    There are no rules in RF2, but we can report same examples about proportions

    of grip between TGM and TBC files:

    S 397 Modern F1 car ( ISI F1 2012 ) ratio between TGM ( sliding grip ) and TBC

    Lat grip: 91.81....% for front
    91,27....% for rear

    As You can see, Lat grip is much more than 80-85% that I previously supposed.

    For vintage F1 car '70 ( S 397 March 711 or Brabham BT 44 )

    ratio Static/sliding grip is 0.733007682 ( more than 0.7 ) due to old tyre features.

    In these cars, Sliding TGM grip and TBC Lat grip:

    ratio 0.955175

    Long TBC grip is 0.969615 of Lat

    Wet TBC grip: Lat and Long are 0.74102564 of Dry values

    For all cars in RF2, anyway, remember that in TBC files grip values Lat/Long

    for Wet compounds are no more dependant from dry values as %,

    but are dependant from RoadGripEffects ( 4 values in TBC ):

    the first value in TBC must be the same as 3° value of GrooveEffects in the corresponding TGM;

    the second value in TBC must be the same as MarbleEffectOnEffectiveload in TGM;

    the third value in TBC must be the same as the 3° value in DampnessEffects of TGM;

    the fourth value in TBC is the same as in TemporaryGriplossfor wetness of tgm,

    but with negative ( - ) number.
     
    PatientRF2fan likes this.
  10. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Thank you for this additional info.
    I admit that I tend to work haphazardly but this doesn't always have reliable results. I long to be repaired to test.
     
  11. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Hello
    Can you tell me what this "WetConductance = (730,730,0,0)" line is for and how these numbers act.
     
  12. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    10,549
    Likes Received:
    5,378
    It's the additional ground conductance for a wet track - the comments pretty much say it all:

    GroundConductance=(1000,0.003,0) // (base, mult, reserved) - thermal contact conductance coefficient to ground = base+(mult*pressure), where pressure is contact pressure and the reserved variable will be used at some later stage.
    WetConductance=(710,730,0,0) // Additional conductance due to (<dampness>, <wetness>,<and reserved for future usage 1>,<reserved 2>)

    (note: 'damp track' is 0 to 50% 'wet', and 'wet track' is 50 to 100% 'wet')
     
  13. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Thank you, great for this explanation.
    I'll get there eventually ;-))
     
  14. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Hi,
    Everything has helped me a lot so far.
    I still have a problem with dry tires (slick)
    on a dry track I have a good balance but on a wet track the AI goes 3 seconds too fast. I touched:
    "WetLatLong = (0.95, 0.95)" but no result
    then "RoadWetnessGripEffect = -7.04" there is a difference but not enough. if I increase too much the AI almost stops in front of the puddles.
    So I tell myself that there must be something else to do or something missing in my files.
    I leave my TBC and TGM forehead for the medium
    A precision, There is only the medium tire for the moment and no rain tires
    TBC
    [COMPOUND]
    Name="Medium"
    FRONT:
    TGM="245-660-13x12 Medium 2013ER-Ms12.tgm"
    DryLatLong=(2.22,2.22)
    WetLatLong=(0.95, 0.95)
    RoadGrooveGripEffect=0.104
    RoadMarbleGripEffect=-0.11
    RoadDampnessGripEffect=-0.10
    RoadWetnessGripEffect=-7.04
    Radius=0.3262
    RadiusRPM=1.751e-6
    Width=0.2415
    Rim=(0.179, 850000, 9000, 3)
    SpringBase=42300
    SpringkPa=1634
    Damper=690
    SpeedEffects=(580,-1.1)
    LoadSensLat=( -6.00e-5, 0.39, 18900)
    LoadSensLong=(-3.00e-5, 0.55, 18900)
    LatPeak=( 0.118, 0.146, 13900)
    LongPeak=(0.118, 0.104, 13900)
    LatCurve="LatSlip"
    BrakingCurve="DecelSlip"
    TractiveCurve="AccelSlip"
    CamberLatLong=(4.6, 0.128, 0.6)
    RollingResistance=182
    PneumaticTrail=4.51e-6
    HeatBasePeak=(0.12, 0.04)
    Heating=(0.61, 0.033)
    Transfer=(0.049, 0.00258, 2.9e-4)
    HeatDistrib=(8.40, 148)
    AirTreadRate=0.002
    WearRate=1.6e-7
    WearGrip1=(0.972,0.986,0.995,0.999,0.998,0.996,0.993,0.990)
    WearGrip2=(0.986,0.978,0.970,0.960,0.950,0.940,0.891,0.763)
    Softness=0.58
    AIGripMult=0.93
    AIHeatRate=1.3e-5
    AIPitThreshold=0.913
    Temperatures=(128, 93) //(95, 78.5)
    OptimumPressure=(80, 0.0122)
    GripTempPress=(2.59, 1.3, 0.771)
    -------------------------------------------------
    TGM
    [Realtime]
    StaticBaseCoefficient=3.14 //ms2.373 x4=2.560 x5 2.615
    SlidingBaseCoefficient=2.09 //ms2.170 x4=2.122 x5*2.10
    AbrasionCurveWLFStartStep=(-5.7,0.50)
    AbrasionVolumePerUnitEnergy=(6.09e-10,5.11e-10,4.29e-10,3.34e-10,2.49e-10,1.92e-10,1.47e-10,1.06e-10,0.76e-10,0.49e-10,0.33e-10,0.28e-10,0.35e-10,0.51e-10,0.75e-10,1.10e-10,1.45e-10,1.95e-10,2.55e-10,3.41e-10,4.27e-10,5.14e-10,6.10e-10)
    DegradationPerWearFraction=(0.9885,0.9998,0.9994,0.9986,0.9978,0.9971,0.9964,0.9957,0.9951,0.9945,0.9940,0.9935,0.9930,0.9925,0.9920,0.9915,0.9910,0.9905,0.9900,0.9895,0.9890,0.9885,0.9880,0.9875,0.9870,0.9865,0.9860,0.9850,0.9838,0.981,0.978,0.863)
    DegradationCurveParameters=(686.15,2550) // 886.15,2550-986-1086
    DegradationPerUnitHistory=(1,0.99,0.971,0.961,0.956,0.953,0.9506,0.9486,0.9467,0.9448,0.9429,0.941,0.9391,0.9372,0.9353,0.9334,0.9315,0.9296,0.9277,0.9258,0.9239,0.922,0.9201,0.9182,0.9162,0.9141,0.9118,0.9092,0.9061,0.9018,0.8968,0.892)
    TemporaryRingMassInertia=(6.536834917655593,0.617868949503828,0.3622028377771681,0.36220283777716944)
    TemporaryRingDamper=(0.035,0.05,0.05,0.133,0.086,0.086)
    TemporaryBristleSpring=(67064,22110,91385)
    TemporaryBristleDamper=(1.39, 1.29, 1.39)
    BaseDamperPerUnitArea=(2340,1881,2574) //2364,1900,2600
    MarbleEffectOnEffectiveLoad=-0.1415
    TerrainWeightOnContactTemperature=0.10
    WLFParameters=(252.15,50,-8.86,51.5)
    StaticRoughnessEffect=-0.20
    GrooveEffects=(0.099,0.075,0.060,0.045)
    DampnessEffects=(-0.28,-0.28,-0.23,-0.18)
    StaticCurve=(176.5, 3.61, 376.5, 1.176, 676.3, 0.61) //178.0, 3.61, 378.0, 1.176, 678.0, 0.61)
    SlidingAdhesionCurve=(-8.9, 0.4, -4.95, 1.63, -1.05, 0.2)
    SlidingMicroDeformationCurve=(-4.95, 0.3, -1.05, 1.8, +2.65, 0.3)
    SlidingMacroDeformationCurve=(-1.05, 0.2, +2.65, 2, +6.1, 0.4)
    RubberPressureSensitivityPower=(-0.21,28500,800000,1)
    IgnitionParameters=(670,0.07,68)
    SizeMultiplier=(1,1)
    TemporaryGripLossForWetness=0.64 //t2012=39-2018=37
    ThermalDepthAtSurface=0.00009
    ThermalDepthBelowSurface=0.00041 //
    BristleLength=0.20
    DampingHeatEnergy=(1.48,0.45,1.0)
    InternalGasHeatTransfer=(14.8,5,0.6) //
    ExternalGasHeatTransfer=(5.5,2.16,0.6) //5.5,2.2
    GroundContactConductance=(2040,0.015) //
    WetConductance=(3400,3130,0,0) //temp)
    TireRadiationEmissivity=1.025
    WetTerrainEffect=(0,1.0,0.95) //t2018(lat,long)usur
    DryTerrainEffect=(0,1.0,1.0) //1,1
    -----
     
  15. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    At a first look at your files, it seems that Rolling Resistance in TBC is too low, while in TGM files are missing the following strings:
    HystereticVerticalDamperPerUnitArea=......
    VerticalDampingMultiplier=....................
    Very important for a correct Tyre working
     
    philrob likes this.
  16. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    Furthermore, in TGM
    TemporaryGripLossForWetness=0.64 is too high.....commonly You should put for a Dry tyre 0.34-0.35
    For Wet tyre 0.15
     
    philrob likes this.
  17. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Thx Nicolas
    I know but it's voluntary
    ----------
    Can you tell me what affects these 2 parameters?
    HystereticVerticalDamperPerUnitArea = ......
    VerticalDampingMultiplier = ........... .........
    ----------
    If I don't set this value, I find driving on a wet track too generous and when the tires slick raise sprays and it doesn't spin, it's unrealistic for my taste.
    ---------------
    Is there no other solution to slow AI down on a wet track with a dry tire?
     
  18. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    The missing parameters in TGM are setting the Rolling resistance of the file, together with hysteresis ( affecting grip ).
    This is the motive for which You are not reaching a good, realistic balance between TGM and TBC
     
  19. Rock

    Rock Registered

    Joined:
    May 28, 2015
    Messages:
    156
    Likes Received:
    18
    Ok.
    Can you give me an example of a standard value for these 2 parameters for testing?
    I can't seem to translate "hysteresis" can you give me an idea of what it means
     
  20. Nicola59

    Nicola59 Registered

    Joined:
    Dec 5, 2014
    Messages:
    101
    Likes Received:
    119
    Hysteresis is the energy of tyre, dissipated by its tread through the movement and attrite on ground ( track ).
    The more a tyre has grip ( proportionally between car weight, engine power, tyre compound and track surface ) , the more that tyre has hysteresis.

    Anyway...some example for You on several tyre models:

    Indy car ( Dallara ):
    BaseDamperPerUnitArea=(2250,2150,2350)
    VerticalDampingMultiplier=(1.25,1.06,0.97,0.96,0.97,1.05,1.25)
    HystereticVerticalDamperPerUnitArea=100

    F1 car
    BaseDamperPerUnitArea=(2364,1900,2600)
    HystereticVerticalDamperPerUnitArea=500
    VerticalDampingMultiplier=(2.9,1.46,0.58,0.62,0.58,1.46,2.9)

    LMP1 car
    BaseDamperPerUnitArea=(3330, 1200, 3874)
    HystereticVerticalDamperPerUnitArea=13000
    VerticalDampingMultiplier=(1.26, 1.01, 1, 1, 1, 1.01, 1.26)

    If I can suggest the way to move, here is:
    first, try to get a good, realistic result with TGM tyres...Tyre must be realistic on both dry and wet conditions, of course.
    Then, as second step, You can work around the TBC file, making the adequate propoprtion of grip ( Lat and Long ) respect to TGMs ( remember that You can test tbc tyres in MOdDev using the option Alt T ). Finally, when tyres TGM and TBC are almost close ( for what it's possible ) in behaviour, try to get with Rolling Resistance in TBC the same max speed available.
    But You have to start with some good tyre models....
    So, depending on the car you are developing, grab some good, ready tyre set ( TGM and TBC ) from any available mod
     

Share This Page