Re: [Slipstream-devel] Rider upper body roll and translation

[Dimitris Papavasiliou]

I think I managed to improve the current rider model.  The setup is as
described in [1] (see fig.8) with the addition of roll freedom for the
upper body.  Basically the rider upper body is connected to the
chassis via a joint that allows roll and yaw with configurable
stiffness and damping.  In addition to that there's a virtual
torsional spring between the upper-body and the front frame (the
handlebars, tripple-tree, forks, front wheel etc.) which simulates the
feedback loop between the front frame and the rider.  In the web
interface the stiffness and damping parameters for the latter can be
found under "Upper body to front frame coupling".

The translational (lateral) degree of freedom for the rider was
removed as it's not very common in published models so it's probably
not important (and little stiffness/damping data is available).
Removing this lateral degree of freedom made the simulation a bit
quicker too, thus making up for the performance loss due to the new
engine model.

The stiffness parameters for the yawing and coupling springs come from
[1].  The damping is given in terms of damping ratios and natural
frequencies and I have no idea how to convert that to damping
coefficients for a two-degree-of-freedom oscillating system such as
this (actually I do have _some_ idea but it involves taking an
introductory course in vibration analysis therefore I fell back to
plan b which is...).  So I guessed at the parameters and I think I
managed to find decent values.  The vehicle is much more stable at any
rate.  Suggestions with better values are welcome.

At some point during the past changes, probably when implementing the
new steering model, something happened and now much larger steering
torques are required to get the motorcycle to turn.  About 10 times
larger than before to get the same effect I'd say.  So if you find you
can't steer, try removing a zero from the first two steering
sensitivity values.

D.

[1] Measurement and identification of the vibration characteristics of
motorcycle riders, V. Cossalter, A. Doria, D. Fabris, M. Maso

On Wed, Jan 11, 2012 at 2:25 PM, Dimitris Papavasiliou
<address@hidden> wrote:
> Hi,
>
>> WRT traction control etc;
>> while I'm sure a preemptive traction control system taking advantage of
>> both acceleration and suspension loadings would be beneficial, we've
>> been riding bikes for years without, so I don't think we should be
>> looking at this as a solution.
>
> Well yes, I did not mean it as a solution just that it might be more
> appropriate than a conventional TC but you're right that it should be
> drivable without it.  Nevertheless I should point out that we probably
> don't ride bikes like this in the real world.  I mean I don't ever
> remember doing clutch-less shifts under full throttle while at a 20
> deg. lean at corner exit :).  It doesn't seem entirely implausible
> that the machine should object by wobbling.  Shakes like these are
> common in race footage.  Still it probably is a simulation issue
> because it's not the initial wobble that's the problem it's the fact
> that it gets worse as if the rider and machine resonate.  I don't
> think that should happen but I'm hoping it's a matter of a
> misconception on my part with regards to the rider model as I explain
> below.
>
>> Thanks for the in-depth discussion of the rider model. From what I
>> understood the rotational stiffness of the upper body is a pure moment
>> connection between the steering head and the body; ignoring any
>> stiffness provided by the torso and lateral force on the bars. If I've
>> got that right, it seems fairly likely that there would be instability.
>> The rider model is quite the complex challenge compared to the rest of
>> the model.
>
> I reread the paper in question and see now that I didn't understand it
> correctly before.  A passage follows:
>
> Thus, in the revised motorcycle–rider model, the
> rider’s upper body has both yaw and roll freedoms
> relative to the main frame. The rider’s arms contribute
> modestly to the steering inertia of the front frame and a
> parallel rotational stiffness and damping element acts
> between the handlebars and the rider’s upper body.
> Following reference [16], the steering inertia contribu-
> tion is fixed at 0.103 kg m2, while the steering stiffness
> and damping parameters are 3.2 N m/rad and
> 0.72 N m s/rad respectively for relaxed riding and
> 60 N m/rad and 1.8 N m s/rad respectively for tense
> riding. From these values, the rider’s upper body
> restraint parameters are estimated as 60 N m/rad and
> 13.5 N m s/rad in yaw and 380 N m/rad and 34 N m s/rad
> in roll for relaxed riding and as 120 N m/rad and 13.5
> N m s/rad in yaw and 760 N m/rad and 34 N m s/rad in
> roll for tense riding.
>
> When I read this initially I was just looking for numbers for
> stiffness and damping for the yawing freedom of the rider's waist so I
> didn't read it carefully.  The way I understand this now is that the
> upper body has yawing freedom at the waist (which is assumed to be
> fixed to the saddle) as if joined with a hinge with a vertical axis.
> It's completely free to yaw.  I does however have some sort of
> coupling with the steering but I'm not sure how to read that "parallel
> rotational stiffness and damping element".  Is it supposed to be a
> virtual torsional spring/damper that doesn't constrain motion at all
> (like a hinge would)  but only applies torque based on the orientation
> difference (and rate of change of it) between the upper body and the
> steering stem?  That seems most likely to me, but then what is the
> axis of this torsional spring?  The vertical or the steering axis?  In
> any case this scenario seems more reasonable as the incorporation of
> yawing in the rider model is introduced to account for feedback at the
> steering so it's reasonable that the stiffness and damping be
> introduced at the torso-steering interface.  I'll play around with it
> once I'm done with the transmission (which is almost complete I think,
> although I didn't notice much difference with respect to the old
> setup).  Let me know if you have a different understanding of the
> passage above.
>
> D.