Hi Everybody,
I've noticed something odd about my vehicle sim, and I was wondering if
anybody else had come across the problem.
I have a vehicle which has 4 tyres. I drive forwards until I get
considerable velocity, then hit hard left on my controller (and hold it
down). This causes my front tyres to rotate some angle theta to the car's
axis. I also keep the accelerator held.
As a result of this slip angle, I generate forces in a lateral direction to
the front tyres (using Pacejka, or whatever method any other method used to
calculate cornering forces). This causes a torque to my car which will
rotate the whole vehicle to the left (-ve torque around the y axis in my
coordinate system).
As the car rotates, the rear wheels also begin to generate forces in their
lateral direction to oppose their motion (+ve torque around the y axis).
OK, now here's the weird bit. The lateral forces at the rear can never be
greater than at the front (due to part of the traction budget being used to
accelerate the car). After initially doing a skid round to the left (as I'd
expect) after a while it begins to spin back partly to the right, before
going back to the left again etc etc.
This seemed odd to me, until I sat down and figured out what was happening.
Currently my tyres are symmetric about the car's COM, and I'm ignoring any
weight transfer due to cornering/braking etc. (ie no suspension). So you
would expect that applying equal forces to all four tyres would have a
resultant torque of zero. However, if the front wheels are turned, then the
magnitude of the torque due to these wheels decreases (since there's a
smaller angle between the tyre's lateral direction and the line from the
tyre to the vehicle's COM). So in the situation above, equal forces to all
four tyres would result in the back end swinging around.
This is what's happening in my sim. As I'm turning, the torque generated by
the rear wheels is gradually increasing until eventually it is greater than
for the front wheels, causing the direction of the overall torque to
reverse. This is behaviour that seems to make sense mathematically, but
which I'm sure doesn't happen in real life (I've never noticed it anyway) or
in other sims I've seen.
Possible reasons I've thought of are different mass distribution, weight
transfer etc, but I can't really see how this would make a difference in all
situations.
I've tried applying the force at the front wheels in the car's lateral
direction rather than the wheel's lateral direction, and this overcomes the
problem (obviously) but it just seems like a bit of a hack to me.
Has anybody else noticed this problem? And if so, what did you do?
Cheers,
Mike.
Hi Mike,
just curious, how are you calculating the velocity of the tyre? It's
important that not only do you use the velocity of the c.g. of the car,
but the additional component which is the angular velocity of the car
vectorial the distance vector from the c.g. to the tyre. This component
is responsible for the yaw dampening which prevents such overshoots. If,
however, the inertia of your car is very large, it may not be sufficient
to stop such motion, and you might want to check into that as well.
-Gregor
Mike Stanley wrote:
Quoted text (55 lines)
> Hi Everybody, > > I've noticed something odd about my vehicle sim, and I was wondering if > anybody else had come across the problem. > > I have a vehicle which has 4 tyres. I drive forwards until I get > considerable velocity, then hit hard left on my controller (and hold it > down). This causes my front tyres to rotate some angle theta to the car's > axis. I also keep the accelerator held. > > As a result of this slip angle, I generate forces in a lateral direction to > the front tyres (using Pacejka, or whatever method any other method used to > calculate cornering forces). This causes a torque to my car which will > rotate the whole vehicle to the left (-ve torque around the y axis in my > coordinate system). > > As the car rotates, the rear wheels also begin to generate forces in their > lateral direction to oppose their motion (+ve torque around the y axis). > > OK, now here's the weird bit. The lateral forces at the rear can never be > greater than at the front (due to part of the traction budget being used to > accelerate the car). After initially doing a skid round to the left (as I'd > expect) after a while it begins to spin back partly to the right, before > going back to the left again etc etc. > > This seemed odd to me, until I sat down and figured out what was happening. > Currently my tyres are symmetric about the car's COM, and I'm ignoring any > weight transfer due to cornering/braking etc. (ie no suspension). So you > would expect that applying equal forces to all four tyres would have a > resultant torque of zero. However, if the front wheels are turned, then the > magnitude of the torque due to these wheels decreases (since there's a > smaller angle between the tyre's lateral direction and the line from the > tyre to the vehicle's COM). So in the situation above, equal forces to all > four tyres would result in the back end swinging around. > > This is what's happening in my sim. As I'm turning, the torque generated by > the rear wheels is gradually increasing until eventually it is greater than > for the front wheels, causing the direction of the overall torque to > reverse. This is behaviour that seems to make sense mathematically, but > which I'm sure doesn't happen in real life (I've never noticed it anyway) or > in other sims I've seen. > > Possible reasons I've thought of are different mass distribution, weight > transfer etc, but I can't really see how this would make a difference in all > situations. > > I've tried applying the force at the front wheels in the car's lateral > direction rather than the wheel's lateral direction, and this overcomes the > problem (obviously) but it just seems like a bit of a hack to me. > > Has anybody else noticed this problem? And if so, what did you do? > > Cheers, > > Mike.Yeah, I'm taking the angular velocity into account when calculating the
tyre's velocity, but one thing I haven't tried yet is changing the inertia
of the car.
I'll play around with it and see what happens, thanks for the suggestion.
"Gregor Veble" <[email hidden]> wrote in message
news:[email hidden]...
Quoted text (12 lines)
> Hi Mike, > > just curious, how are you calculating the velocity of the tyre? It's > important that not only do you use the velocity of the c.g. of the car, > but the additional component which is the angular velocity of the car > vectorial the distance vector from the c.g. to the tyre. This component > is responsible for the yaw dampening which prevents such overshoots. If, > however, the inertia of your car is very large, it may not be sufficient > to stop such motion, and you might want to check into that as well. > > -Gregor >Mike,
If you're talking about getting a bit of a wobble when you crank the wheel
hard, that's right. This is why real drivers turn the wheel smoothly and never
try to jerk it around. Keep the force going in the tire's direction, rather
than the body's direction. As Gregor pointed out, changing inertia values
(around the yaw axis) will change exactly how this works.
Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com
Unless you're a rally driver, and you deliberately yank it in one, then
the other direction so that it overshoots sideways (a pendulum turn, or
so I hear) :).
-Gregor
"J. Todd Wasson" wrote:
>
> If you're talking about getting a bit of a wobble when you crank the wheel
> hard, that's right. This is why real drivers turn the wheel smoothly and never
> try to jerk it around.
"J. Todd Wasson" <[email hidden]> wrote in message
news:[email hidden]...
> Mike,
>
> If you're talking about getting a bit of a wobble when you crank thewheel
> hard, that's right.
Yep, finally seen it in a game. Viper racing in simulation mode does the
same thing when doing a hard turn at high speed.
I'm happy now, must be doing something right :)
Viper really is a good game but under-rated, it could easily be a great sim
if it had real tracks, tire wear,ect.. I really like the way some of the
physics is done in Viper and in certain area's it feels better then GPL and
N4. Nascar Heat has been dumbed down or compaired to Viper.
I'm not sure why but Viper racing is the only game I feel comfortable
driving with 2 hands. I can't drive with 2 hands with GPL, N4, it just
doesn't feel right or comfortable.
Driving the GTS/R is lots of fun. Now if someone could make some real tracks
for Viper racing that would be great!
"Mike Stanley" <[email hidden]> wrote in message
news:W_6f7.5094$[email hidden]...
