[ODE] Forward-correcting force

[Jean-Sebastien Guay]

Hello,

I'm having trouble refining the idea of forward-correcting force that Jon Watte
gave me the first time I asked about modeling the movement of a submarine
underwater.

What I would like is that if the sub currently has a nonzero linear velocity,
but that I turn it a bit so that its direction vector is no longer lined up
with the linear velocity's direction, the linear velocity should come back
gradually (quite quickly in fact) to line back up onto the sub's direction
vector.

It isn't that hard, normally I would think a quick dot product and then a force
in relation to that would work. But I'm still having trouble with it. Here's
what I have:

    Vector right = m_Direction .Cross(m_Up);

    // '*' between 2 vectors is dot product in (GetLinearVelocity() * right)
    Vector fcf = m_Direction *
        (GetLinearVelocity() * right) * -1.0f * m_Mass;
    GetPhysicsObject()->AddForce(fcf);

The problem with this is that it does bring back the orientation to the
direction (not fast enough yet, but I can tweak that later), but it introduces
too much new speed. My sub is supposed to have a given max speed (m_MaxSpeed)
and after a few turns with this force enabled, I'm going really a lot faster
than that.

Just for reference, my acceleration function goes roughly like this:

    if (GetLinearVelocity().GetLength() < GetMaxSpeed())
    {
        Vector force = m_Direction * m_Mass * m_Acceleration;
        GetPhysicsObject()->AddForce(force);
    }

Perhaps there's a better way of bounding the linear velocity of an object?

Can anyone point me in the right direction please?

--
______________________________________________________
Jean-Sebastien Guay     jean-sebastien.guay at polymtl.ca
                         http://whitestar02.webhop.org