Advertisement

Race Car Physics

Started by September 11, 2000 04:43 PM
13 comments, last by Grasshopper 24 years, 3 months ago
Pretty good explaination ahw, and I agree, this situation really begs for some pictures. Its hard to describe it otherwise... here''s my stab at it.

First off, skidding is simply where the tires can no longer supply enough frictional force to counteract the forces acting on your tire. They no longer have static friction with the ground, but instead kinematic friction, which is a fancy way of saying that they just start dragging like any ol object. A rotating wheel usually has static friction, because the contact point that is on the ground is not actually moving.. skidding changes this.

Once your tires start skidding, they act like a regular sliding object would, ie they continue in whatever direction they were going, with friction acting in the opposite direction of that. Eventually this friction force slows the tire down enough that the sideforce drops below the limit of kinematic friction, and your car regains its footing.

So there''s two types of skid that you want to model : when the front (turning) wheels skid, and when the rear wheels skid. They both work exactly the same, but you''ll probably want to seperate them, because the car will handle differently depending on which wheels are skidding.

In both cases, you determine if the car is skidding out by the amount of force being exerted in a non-vertical direction on the tire.

so if the force we''re exerting (on the rear tires from accelerating, the front tires from turning, etc) is greater than this maxforce, our tires will go into a skid until they go under the threshhold -- now, an important point is that the value where your tires regain their grip is at much less force than you could originally put on them. Try it yourself. Take a book and put it on an inclined board. Raise the board up until the book starts sliding. You''ll have to lower the board much lower to stop the book from moving farther down. This is the difference between our static and kinematic (moving) friction

So, lets apply this to our car.

You''re cruising along at 100kph (yeah I don''t use metric normally either, but I learned physics in the metric system). You start a sharp turn to the right. You turn the front wheels, lets say, 15degrees, which is enough to bump your sideways force over the limit we looked at earlier.

Now, the ground will apply a friction force on your tires, opposing the motion of the car straight forward.

Once again, you can check if a tire is past its adhesion limit by checking the sideways force limit :

maxforce on tire = weight in kg being put on this tire * 9.8m/s^2 * static frictional coeffeicient

When you''re actually in a skid, you use the same formula, except to use the kinematic frictional coeffeicient instead of static. (just a lower number)

So, as this force is applied, your car slows down forward. You probably also want to model partial grip, where your tires, even though they are skidding, still accelerate you in the direction you''re pointing, just at a very reduced rate.

Another common thing is for the rear end to skid out. Once again we just use the same checks, see if there is a non-vertical force greater than our limit, if so, the rear tires will no longer track, but will continue in the same direction they were before, their speed being reduced every timeslice by the frictional force until it drops low enough to reduce the sideways force.

An example of this would be when you are moving, and just after starting your turn, either punch the accelerator or pull the emergency break, putting the rear wheels into a skid. The rear wheels and thus the rear end of the car continue on their previous path, while the front wheels track the curve, resulting in a very familiar slide-out, where the back of the car is moving sideways compared to the direction the car is pointing.

To get this to work right, you will have to model 3 things :

The front tires
The rear tires
The car body

If the front or rear tires go into a skid, the differences in the location of your two ends defines the angle of the car. This gets a bit more complicated also, because then you also need to model the force between the components -- for instance, the rear tires are in a skid, going sideways in comparison to the front tires. The front tires will provide some additional frictional force to stop the skid, since they are attached to the rear tires through the body, and the relative locations of the tires and body cannot change (without your car getting wrecked, of course. )


There''s tons of other things to try modelling in your car also, such as weight distribution, roll, springs, etc. But I think realistic skids and power is more than enough to chew on at start.

-- Remnant
- remnant@mailandnews.com


There''s a bunch of physics articles over at Gamasutra that may help you out: http://www.gamasutra.com/features/index_programming.htm



Advertisement
wow, someone got my explanations ?

Agreed, you need to separate front and rear wheels, as the front ones are those that turn, and its the resistance of the tires that allow you to turn ... unless you turn too thigthly in which case they start sliding.

One thing I''d like to know, is the inlfuence of the fact that a set of wheel can be following, of actually pushing (front wheel motor, rear wheel, 4x4 drive).
I think also, there is an intersting article in Gamasutra about the transition between sliding and friction, which you need to use if you want to simulate the things like ABS (stop breaking to get friction again).

As well, one interesting thing to do when you get far enough into this kind of physics is to take into account the tilting effect And why not make the car tilt over (is cases of extreme skidding).

but I guess as well, if you get realistic sliding, you''d have a pretty cool game already... I don''t know a lot of game that id just *that* in a nice, playable and yet reaslistic way...

good luck
-----------------------------Sancte Isidore ora pro nobis !
Here is the URL of the phor article: Physics of Racing Series
http://members.home.net/rck/phor/
That''s a very nice article explaining all the physics of racing!

It''s good to see that, even though I was discovering my explanations about racing by applying my physics knowledge as I went, I was more-or-less accurate.

-- Remnant
- remnant@mailandnews.com

This topic is closed to new replies.

Advertisement