0
\$\begingroup\$

When an object is controlled by a CharacterController, it is easy to make it climb ramps with a certain slope, or climb steps with a certain height.

When the object is controlled by a RigidBody instead, it can still climb ramps naturally - I just use AddForce and it walks on ramps just like on any other surface. I do not have to add any special logic for a ramp. However, it does not climb stairs. Is there a natural "physical" way to make the RigidBody climb stairs, without explicitly checking that there are stairs ahead?

\$\endgroup\$
2
  • 1
    \$\begingroup\$ What shape is your collider? Does it have a rounded edge that can lift it up and over the step, or a hard corner that will snag? \$\endgroup\$
    – DMGregory
    May 14 '20 at 11:46
  • \$\begingroup\$ @DMGregory It is a capsule. \$\endgroup\$ May 14 '20 at 13:21
1
\$\begingroup\$

In order to lift a capsule up the stair's riser, the normal force exerted by the stair corner against the rounded bottom must meet or exceed the force of gravity pulling it down.

Given a particular capsule radius \$r\$ and riser height \$h\$, both sitting on the same floor plane, the corner hits the capsule at a point relative to its center of curvature:

$$ \vec p_{contact} = \left( \sqrt{r^2 - h^2}, h - r\right)$$

Which means the normal force will act in the direction given by the unit normal:

$$\vec n = \frac {-1} r \left( \sqrt{r^2 - h^2}, h - r\right)$$

To lift off the ground, the vertical component of this force has to equal or exceed the force of gravity, given by an acceleration \$a_g < 0\$:

$$\begin{align} f \cdot \frac {h - r} r &= m \cdot a_g\\ f &=\frac { m \cdot a_g \cdot r} {h - r} \end{align}$$

This will then try to push your capsule backward away from the stair by the horizontal component of the force:

$$f_{horizontal} = \frac { m \cdot a_g} {h - r} \sqrt{r^2 - h^2}$$

So you'll need to drive your capsule into the stair with a force greater than or equal to this amount to achieve lifting. Plus some extra to overcome losses to friction.

\$\endgroup\$
1
  • \$\begingroup\$ Very interesting, thanks! So the reason that my capsule did not climb the stais was just that the force was too small. \$\endgroup\$ May 19 '20 at 8:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.