[ODE] Questions about wheel slip and damping

[Michael Rauh]

Hi Yizhen, I can't answer in the same length as your questions ;), but
you should just try two things:

- set the joint motors of the unpowered wheels to the same velocity as
the powered ones (but leave their torque at zero)

- for suspension, just set the cfm/erp of the hinge-2 as the docs
state it ("how to use erp and cfm")

That should help. 

Greets
Michael


On Mon, 11 Oct 2004 12:10:33 -0700 (PDT), Yizhen Zhang
<zhangyizhen at yahoo.com> wrote:
> Dear all,
> 
> I am pretty new to the ODE community. I am just trying
> to implement some basic vehicle dynamics on a simple
> car model built in Webots. The current model only
> contains four cylinder wheels and one box body. The
> two front wheels are driving wheels which I apply
> torque on, and the two rear wheels were motor off.
> 
> First I noticed that when the car finished
> acceleration and entered steady state (moving forward
> at constant speed) stage, the following relationship
> seems to hold:
> v = (wf + wr) * r / 2                 (1)
> where v is linear speed of the car, wf and wr are
> angular speeds of the front and rear wheels, and r is
> the wheel radius.
> 
> The other thing I noticed is that there is always a
> difference between front and real wheel angular
> speeds. And this difference seems to depend on wheel
> radius. Actually when the wheel radius is too small
> (say in centimeters), the rear wheels do not rotate at
> all, and they look locked and just slide forward, i.e.
> wr = 0. But equation (1) still holds, so I get v = wf
> * r /2, i.e. the front wheels slip at 50%. When I
> increase the wheel radius (say, to > 0.2m), the rear
> wheels start to turn, but I always get wf > wr, and
> the following seems to hold:
> (wf - wr) * r * r = constant        (2)
> So the rear wheels still skid as if some brake force
> were applied to them. I tried to change friction
> coefficients and FDS, but (1) and (2) always hold for
> steady state. I wonder why it is implemented like this
> in ODE, and I can't believe this is the case in real
> cars. I hope there is other ways to control the wheel
> slip percentage instead of changing wheel radius.
> 
> On the other hand, I defined a simple slider joint
> between the box body and the car body (the part move
> with wheels), so that the box could move up and down
> relative to the car. I just want to implement a very
> simple spring and damping system to simulate the
> suspension. So I added opposite spring forces to both
> bodies, which are proportionate to the vertical
> distance between the two. However I observed damping
> effect even I only added spring forces. I can
> understand if ODE developers added damping here to
> stablize the simulation. But I wonder how does ODE
> decide how much damping to add here. It seems to be
> small, around 0.01, but changes with the spring
> constant and maybe some other factors.
> 
> Sorry for the message length and I appreciate your
> attention and help.
> 
> Best regards,
> Yizhen
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode
>