Feature: Utility Options #38
Conversation
mbpeterson70
commented
Nov 23, 2025
mbpeterson70
commented
- add option to use a static tf that is published in /tf in data_params
- allow skipping non-overlapping submaps and self loop closures
- make the offline rpgo result plot nicer
| if not np.isnan(robots_nearby_mat[i, j]): | ||
| relative_yaw_angle = transform_to_xyzrpy(T_ij)[5] | ||
| submap_yaw_diff_mat[i, j] = np.abs(np.rad2deg(relative_yaw_angle)) | ||
| if submap_distance > sm_io.skip_distance: |
There was a problem hiding this comment.
Consider moving this after the submap_descriptor calculations, so that submap descriptor similarity is still saved?
There was a problem hiding this comment.
Yeah, I like that, sounds good.
roman/params/data_params.py
Outdated
| ) | ||
| # by default looks for a static tf, but if the user wants to reference a tf that is | ||
| # theoretically static, but is published under /tf, then 'try_non_static_tf' can be set. | ||
| if param_dict.get('try_non_static_tf', False): |
There was a problem hiding this comment.
The naming 'try_non_static_tf' makes it sound like it would look under both /tf_static and /tf as a fallback. Maybe just get rid of the param and use a try/catch fallback instead?
There was a problem hiding this comment.
Haha that's actually what I was doing before but I realized that PoseData.from_bag_tf is basically a super-set of any_static_tf_from_bag, so if the user is okay with the fact that they might be putting in a transform in their tree that shows up in /tf then we might as well use the PoseData.from_bag_tf. What do you think about renaming from try_non_static_tf to include_non_static_tf ?
* update readme with static_tf changes
2 participants
