Output point cloud shifted for dynamic scenes if scene bounding box changes

[han16nah]

The problem
I am simulating a dynamic scene with rigid motions, where a branch is rotating about a point at the bottom of the stem. By this rotation, the bounding box of the scene also changes. If I visualize the simulated point cloud together with the static base mesh, I would expect that the rotation center is at the same position. However, the simulated point cloud is shifted and the rotation center is not at the same location as the one for the static base mesh (see screenshot below).

I also have a static scene for this same case. Both simulations over static and dynamic scenes give the same bounding box and internal HELIOS++ scene shift (as shown in the log).

CRS bounding box (by vertices): Min: dvec3(-0.044812, -0.223585, -0.002277), Max: dvec3(0.163092, 0.126298, 0.448723)
Shift: dvec3(0.059140, -0.048644, 0.223223)
# vertices to translate: 6096
Actual bounding box (by vertices): Min: dvec3(-0.103952, -0.174941, -0.225500), Max: dvec3(0.103952, 0.174941, 0.225500)

I therefore assumed that the problem might be the change of the bounding box as the scene geometry changes, resulting in incorrect internal shift/translation. Hence, I added a large static "box" around the scene. With this workaround, the shift is gone and the simulated "4D" point cloud is as expected.

XML files to reproduce

helios_data.zip

Environment:

• OS: Windows 11
• HELIOS++/pyhelios version: v2.0.2

Screenshot

• Base mesh (that undergoes rigid motion)
• Wrong simulation output in RED
• Expected simulation output (when adding box around the dynamic scene) in BLUE

Note: Scene only undergoes rotation, the rotation center is the bottom of the branch