Walk on surface rotation issue

using depth pipet for walking on a mesh,
how does the rotation map to clamp the unnecesary spin ?
mesh along.v4p (33.8 KB)

a similar math issue comes up when sampling an sdf surface for its normal while moving.

  • using closestpoint/projectpoint (transform) results in a type of rotation skipping.
    could that be solved with quaternions or some ‘lazy’ angles remapping ?
  • using closestpoint/projectpoint (3d buffer),
    i tried to define a plane from three points
    but its unclear how? to retrieve the transform for the walking mesh.
    (nearest node seems paralelepiped but that applies scale also and
    decomposing paralelepiped fails at some of the projected angles)

mesh on (79.1 KB)