-
Type:
Task
-
Status: Open (View Workflow)
-
Priority:
Normal
-
Resolution: Unresolved
-
Component/s: ets_fiberalloc
-
Labels:None
When trying to calculate trajectory collisions, either using
CollisionSimulator or CollisionSimulator2 from ics.cobraOps, I get NaN values for the cobra position angles if the final positions are the same as the home position. The NaN values break the subsequent testing of the trajectory collisions aroundhttps://github.com/Subaru-PFS/ics_cobraOps/blob/81d6a313d3eea21fcccfd970596617668c0f8edb/python/ics/cobraOps/TrajectoryGroup.py#L131
The issue happens when I want to test an existing pfsDesign file which has some unassigned fibers. In this case, the home position is given for each fiber but it might be slightly different (up to about 1e-7) from the actual home position because of rounding errors when the pfsDesign file is saved with 32 bit precision.