Tire friction model overreacts to pressure

There seems to be a problem with the wheel friction model that causes unrealistic amounts of grip.

I posted a picture that should help me explain this better. One thing that really messes up a racing game is when the traction model gets messed up when the car is running around a banked turn.

To make things worse, my game uses many different road types, that tend to put the car under a lot of pressure.

For example, a loop (see screenshot)

In this particular example, the car rapidly accelerates to 200 MPH + by time it ends the loop.

Does anyone have any suggestions on how I could reduce the impact on grip when the car is under pressure?

Hmm… I tried doting the suspension values… It’s just not as responsive.

Is there a function that finds the gforce of an object?

Couldn’t you simply calculate the g-force yourself?

example:

a = ∆V/∆t

∆V = 100 km/hr (also 27.7 m/s (V changes from 0 to 100 km/r over ∆t)

a = 27.7/2 = 13.8 m/s² (this change happens in 2 seconds, ∆t = 2)

We know that g = 9.8 m/s²

The g-force is then 13.8/9.8 = 1.4 g

Obviously, working with skew lines will introduce more forces, so it wouldn’t be as simple as this. Perhaps check up on centrifugal forces?

Hmm… I have no idea how to use the delta ∆, or whatever they are. The worst I can handle is trigonometry. :stuck_out_tongue:

I’m not sure… :confused:

So far, Ive put two things together in my mind: Velocity and Spin

If the car velocity along local z is fast, and the local y is slow and the car is spinning, those three conditions guarantee gforce exists.

But I think that only covers the trajectory… I also need impact gforce.

I’ll have to give it some time while this is cooking in my brain.

Delta just means “a change in” something. So ∆t means “a change in” time, in other words, time1 - time 2.

Ok that should help me understand this better…
Thanks

The wheel collider has a rather unrealistic nature. Impossible grip levels, unexpected drifting, very high acceleration occurs from time to time. For instance changing gravity increases the cars top speed, which it should not. I guess what you experience is similar. Some people prefer to make their own wheel collider, with their own physics. If you want a realistic car, I guess it is the only way for now.

Tig: That’s exactly it. The game engine I used to use did not have this problem and still used the PhysX car. There must have been a special way to reduce this effect. I’m still looking into it now.

I’ve got a buddy who is creating a new game engine with PhysX in it, so I’ll ask him if he has the same problem.

Just curious; what physX sdk version does Unity use?

Is unity 3.0 at all related to physX 3.0?
http://developer.nvidia.com/physx-sdk-V30

Well… I got my problem solved. I basically made a G-force meter and used it to multiply the traction. I also went through and plotted some Math.Asin’s and that smoothed out the bad reaction to landing jumps.

Great. Did you use the existing wheel collider or made your own?