PDA

View Full Version : Solved: Vehicle centrifugal force problem



Xander112
05-26-2010, 02:36 PM
Hi,

I'm having some problems with a vehicle I have made. Everything is working fine except the centrifugal force while in corners. It actually works but its the wrong way around. When I steer hard left the car doesn't lean right but it leans in left. This is the wrong way around.

The same is happening for breaking. The vehicles mass moves backwards instead of continuing forward while breaking.

I tried a lot of variables in the car's code but nothing seems to change that.

Does anybody know I can fix this?

Saishy
05-26-2010, 05:40 PM
You mean Centripetal force, there's no such thing as centrifugal force.

It just seems your cart is changing the whole vector's direction when changing directions, you should try to change it more smoothly.
Like, when breaking, try to slowly reduce the forward movement instead of just reverting the velocity vector.


http://imgs.xkcd.com/comics/centrifugal_force.png

MonsOlympus
05-27-2010, 03:40 AM
Wait no this sounds correct the roll of the vehicle should be that of the turning direction. Youre asking about lateral and roll dampening removal?

Saishy
05-27-2010, 09:45 AM
Wait no this sounds correct the roll of the vehicle should be that of the turning direction. Youre asking about lateral and roll dampening removal?
No, I'm pretty sure he is saying the cart is not keeping momentum, when you turn left at high speed you should lean right because the inercia is trying to keep you from turning.

Ok, I'm horrible with explanations.

MonsOlympus
05-27-2010, 12:04 PM
'You' being the person driving the vehicle with the GForces applied, the vehicle chassis or vehicle body? I know what he's getting at, like a spinning top, vehicles arnt fully rigid unless we are talking pre-1979 (I think it is) suspension where we had rigid rears or 4wd/monster truck suspensions. Full independent does some weird things, certain power to sets of wheels (FWD/RWD/A-4WD) causes different effects on turning (under/oversteer) which is lateral slide and camber has something to do with that also. Im pretty sure Roll dampening stops the vehicle going into a barrel roll after a jackknife but the keep upright constraints make the vehicle fall like a feather.

Blade[UG]
05-27-2010, 03:55 PM
Centrifugal force is often confused with centripetal force. Centrifugal force is most commonly introduced as a force associated with describing motion in a non-inertial reference frame, and referred to as a fictitious or inertial force (a description that must be understood as a technical usage of these words that means only that the force is not present in a stationary or inertial frame).


- Wiki

see, there is such a thing.

Xander112
05-27-2010, 05:29 PM
hey, yeh centrifugal force is the opposite of centripetal force.
http://en.wikipedia.org/wiki/Centrifugal_force Its written on wikipedia, so its gotta be an real thing :P
Anyway.
The problem is that somehow when I turn left the right wheels lift. But the left wheels should lift because the weight wants to go straight forwards, when going left thats the right side of the car:
http://regentsprep.org/regents/physics/phys06/bcentrif/casette.gif

I think it mush be a suspension problem, because im extending the UTVehicle and using a lot of settings the same as the Scorpion. The default properties im using:


UprightLiftStrength=280.0
UprightTime=1.25
UprightTorqueStrength=500.0
bCanFlip=true
bStickDeflectionThrottle=true
HeavySuspensionShiftPercent=0.75f;

bFollowLookDir=true
bIsConsoleTurning=true
bUsingLookSteer=true
bStickDeflectionThrottle=true

bLookSteerOnNormalControls=true
bLookSteerOnSimpleControls=true
LookSteerSensitivity=2.2
LookSteerDamping=0.07
ConsoleSteerScale=1.1
DeflectionReverseThresh=-0.3

Begin Object Class=UDKVehicleSimCar Name=SimObject
WheelSuspensionStiffness=90
WheelSuspensionDamping=6
WheelSuspensionBias=0

MaxBrakeTorque=0
StopThreshold=100

MaxSteerAngleCurve=(Points=((InVal=0,OutVal=45),(I nVal=600.0,OutVal=15.0),(InVal=1100.0,OutVal=10.0) ,(InVal=1300.0,OutVal=6.0),(InVal=1600.0,OutVal=1. 0)))
SteerSpeed=110

LSDFactor=0.0
TorqueVSpeedCurve=(Points=((InVal=-600.0,OutVal=0.0),(InVal=-300.0,OutVal=80.0),(InVal=0.0,OutVal=130.0),(InVal =950.0,OutVal=130.0),(InVal=1050.0,OutVal=10.0),(I nVal=1150.0,OutVal=0.0)))
EngineRPMCurve=(Points=((InVal=-500.0,OutVal=2500.0),(InVal=0.0,OutVal=500.0),(InV al=549.0,OutVal=3500.0),(InVal=550.0,OutVal=1000.0 ),(InVal=849.0,OutVal=4500.0),(InVal=850.0,OutVal= 1500.0),(InVal=1100.0,OutVal=5000.0)))
EngineBrakeFactor=0.025
ThrottleSpeed=0.2
WheelInertia=0.2
NumWheelsForFullSteering=3
SteeringReductionFactor=0.0
SteeringReductionMinSpeed=1100.0
SteeringReductionSpeed=1400.0
bAutoHandbrake=false
bClampedFrictionModel=true
FrontalCollisionGripFactor=0.18
ConsoleHardTurnGripFactor=1.0
HardTurnMotorTorque=0.7

SpeedBasedTurnDamping=20.0
AirControlTurnTorque=40.0
InAirUprightMaxTorque=15.0
InAirUprightTorqueFactor=-30.0

// Longitudinal tire model based on 10% slip ratio peak
WheelLongExtremumSlip=0.1
WheelLongExtremumValue=1.0
WheelLongAsymptoteSlip=2.0
WheelLongAsymptoteValue=0.6

// Lateral tire model based on slip angle (radians)
WheelLatExtremumSlip=0.35 // 20 degrees
WheelLatExtremumValue=0.9
WheelLatAsymptoteSlip=1.4 // 80 degrees
WheelLatAsymptoteValue=0.9

bAutoDrive=false
AutoDriveSteer=0.3
End Object
SimObj=SimObject
Components.Add(SimObject)

I think i just need to adjust some setting here but i cant find it.

Saishy
05-27-2010, 06:39 PM
Centrifugal force is fictional, it's a common mistake to think there is a force from inside the rotation that pushes you away from it. But that force is not real, as the wiki says.

I can't help you with actual code, since I'm just a game designer.


The problem is that somehow when I turn left the right wheels lift. But the left wheels should lift because the weight wants to go straight forwards.

No, the right wheel should lift. It's acting correctly.
Technically, no wheels should lift if the car is on a plane, but if your left wheels lift, the car will either flip or fly away.

Blade[UG]
05-27-2010, 07:05 PM
"referred to as a fictitious or inertial force (a description that must be understood as a technical usage of these words that means only that the force is not present in a stationary or inertial frame)."

not fictitious as in "does not exist", but "does not exist in a stationary frame"

MonsOlympus
05-28-2010, 12:26 AM
Yeah thats lateral slide brought on by momentum, anything fixed to the car obviously isnt going to move until the car itself gains enough momentum to start sliding. What happens usually is a car will slide sideways and friction of the wheels goin side ways cause a car to roll, because they are skinnier then they are long, cases which is why cars with wider wheel bases have better handling. Like a chevy corvette c7 or a ford gt40, hummers another example though 4/awd. This is what Ive always called a jackknife, pretty sure thats a slang for it cause the car does a j shape then cuts in.

What you'll find is the car will roll into the corner, this causes the force to be directly down at 90 degree's to the body but a harder corner means the body roll vs the chassis angle is dramatic causing this force towards the outside of a corner. Alot of things will play a part in this:

Firstly is the type of suspension, as I mention rigid rears work differently to full independent. Full independent is automatically going to give you full body roll with less tweakin (scewing) of the chassis in hard cornering.

Secondly is the stiffness of the suspension, if its soft and has alot ride (distance between bound and rebound max) then it'll hug corners like a dream and any chances of lateral slide does kinda go out the window. On the interior everything not attached will have G forces but anything attached will just move in the lateral slide direction once friction (traction) gives in.

The most important one is the keep upright constraint, this is an inertial dampener and is complete science fiction. Real cars use AWD and complex traction control systems to make sure power gets down and in corners the wheels turn the right speeds to avoid loss of control. Basically a keep upright pushes the body back up 90 degree's to the chassis (it might even be world relative) which makes it a kind of anti-gravity :cool:

Edit: I might as well list it here, there are alot of driving assists for UDK cars:
Auto-Brake assist - in cornering and when not throttling. (engine brake)
Steering assist - auto correction when not corning the vehicle will straighten back up.
Anti-Roll assist - the inertial dampener I mentioned, it keeps the vehicle upright obviously and makes it fall like a feather.
Slide dampening (Super Traction Control) - stops the vehicle from going into long uncontrolled slides on the driving plane.
Theres more, basically cars in UDK are like super remote control cars :p

Xander112
05-28-2010, 02:10 AM
According to my brother (car engineer) the problem could be that the breaking force and the centripetal or centrifugal force are applied to a point higher then the ground. Then that behaviour would happen. But when I look at the model im using the root bone is on the floor.
Is there a variable for a offset?

Edit: Ok silly me... pls dont flame. The only thing I had to change was my COMOffset, it was placed under the car causing that behaviour.....
Sorry if i wasted your time,
But thanks for the support.

Xander

MonsOlympus
05-28-2010, 02:57 AM
Im pretty sure the COM is more fictional then centrifugal force. Since, the body has weight, so does the chassis and most importantly the engine. So if you want accuracy you'll need an engine mass and location, this way you'll get the right body pitch you were looking for when braking :cool:

mesavage
05-28-2010, 09:33 AM
i wonder what a certifuge does and why its centrifugation is so important to our scientific society *roflmao*

Saishy
05-28-2010, 09:41 AM
i wonder what a certifuge does and why its centrifugation is so important to our scientific society *roflmao*
I also wonder what a "certifuge" does xD

I think you're being sarcastic, but a centrifuge is used to separate substances based on their density acording to the sedimentation principle.

They are also evil, since they use a force that shouldn't be there.
Fear and respect them.

mesavage
05-29-2010, 03:30 AM
I also wonder what a "certifuge" does xD

I think you're being sarcastic, but a centrifuge is used to separate substances based on their density acording to the sedimentation principle.

They are also evil, since they use a force that shouldn't be there.
Fear and respect them.

Ye, true, I was being sarcastic, taking a stab at the whole "there is no such thing as centrifugal force". I mentioned centrifuge and centrifugation, since centrifugation is the process that a centrifuge does, that involves the use of the centrifugal force for the separation of mixtures.

It defies newtonian law .... muwhahahaha *evil grin*

http://en.wikipedia.org/wiki/Lamm_equation

TheOneandOnly
05-29-2010, 05:52 PM
Actually if you think about it it's fairly clear that it doesn't. If the force acted outwards from the radius the substances would just mix together at the far end. The fact that the substances have different densities (different masses, by extension) means that their forward velocity is affected by a different amount by centripetal force, and they settle at different distances from the centre of rotation.

AlbinoWarrior
07-28-2010, 10:40 AM
Centrifugal force does not exist. Anywhere or Any-when. There is no force that pushes out from the center of the turn. There is only the objects forward momentum.

Blade[UG]
07-28-2010, 10:45 AM
Centrifugal force is often confused with centripetal force. Centrifugal force is most commonly introduced as a force associated with describing motion in a non-inertial reference frame, and referred to as a fictitious or inertial force (a description that must be understood as a technical usage of these words that means only that the force is not present in a stationary or inertial frame).

yes, yes there is a centrifugal force. and "fictitious" here does not mean "does not exist".