IrisZo: Better Handle Cases when the Center of the Ellipsoid is In-Collision

[cohnt]

Closes #22714

Note that we make two changes here:

1. We throw an error if the user provides a seed point that is in-collision (or violates one of the additional constraints the user has provided)
2. If after the first iteration, the center of the ellipsoid moves to a configuration that's in-collision, we return the polytope from the previous iteration, and warn the user that this has occurred.

---

This change is 

[cohnt]

+@wernerpe for feature review, please!

Reviewable status: 2 unresolved discussions, LGTM missing from assignee wernerpe, needs platform reviewer assigned, needs at least two assigned reviewers, commits need curation (https://drake.mit.edu/reviewable.html#curated-commits) (waiting on @cohnt)

---

-- commits line 1 at r1:
working.. will set final commit message later

---

planning/iris/iris_zo.cc line 675 at r1 (raw file):

      break;
    }
    if (!checker.CheckConfigCollisionFree(

A note to the eventual platform reviewer -- although we know this is possible in theory, we haven't been able to reproduce it in practice. So we'd like to include the solution, even though we don't yet have a test case that triggers this code path.

Note that besides the different triggering conditions, the logic is exactly the same as the case where we set options.max_iterations to 1.

[wernerpe]

Reviewable status: 2 unresolved discussions, needs platform reviewer assigned, needs at least two assigned reviewers, commits need curation (https://drake.mit.edu/reviewable.html#curated-commits) (waiting on @cohnt)

[cohnt]

+@rpoyner-tri for platform review, please!