RobotLocomotion / drake

Model-based design and verification for robotics.
https://drake.mit.edu
Other
3.25k stars 1.26k forks source link

Update multibody call sites to use RigidTransform (part 3). #11193

Closed amcastro-tri closed 5 years ago

amcastro-tri commented 5 years ago

Overarching issue: https://github.com/RobotLocomotion/drake/issues/9865

Continuation of the work started in #10959 and #11064.

As with #11064, feel free to assign your name and address as many of these you'd like.

A brute force search of Isometry3 within drake/multibody and pydrake/multibody revels the following 109 occurrences of Isometry3 across 25 files (number counts how many times in that file):

amcastro-tri commented 5 years ago

This has gone a bit stale. I believe @mitiguy is working on some of this, could you confirm @mitiguy? I'll ask for more volunteers in the dynamics channel.

mitiguy commented 5 years ago

Yes @amcastro-tri I am on this. I got side-tracked slightly with numerical integrator test.

mitiguy commented 5 years ago

Monday update: Work to restart after onsite. Started PR# 11414 to remove some additional Isometry and add required friend operator * to SpatialVector class.

mitiguy commented 5 years ago

Progress update Monday May 20th -- background information only All progress last week was on updating the Jacobian interface (Issue in progress #10155), none on this issue. Merged PR#11414 12 days ago which checked off the following boxes in this issue: multibody/benchmarks/kuka_iiwa_robot/drake_kuka_iiwa_robot.h: 1 @mitiguy multibody/benchmarks/kuka_iiwa_robot/make_kuka_iiwa_model.cc: 1 @mitiguy multibody/plant/test/frame_kinematics_test.cc: 7 @mitiguy multibody/plant/test/kuka_iiwa_model_tests.h: 1 @mitiguy multibody/plant/test/multibody_plant_jacobians_test.cc: 1 @mitiguy multibody/plant/test/multibody_plant_test.cc: 11 @mitiguy

amcastro-tri commented 5 years ago

11414 landed. Two drake/multibody bullets remaining, @mitiguy is it a good assumption that you'll address them?

@SeanCurtis-TRI, regarding the last bullet mentioning an additional GeometryInstance ctor, is that a sensible solution?

mitiguy commented 5 years ago

@amcastro-tri Yes - I'll address them this week.

amcastro-tri commented 5 years ago

Excellent! we are only one bullet away from resolving this one! Thanks to everyone who contributed.

jwnimmer-tri commented 5 years ago

FWIW I don't think we should do anything about the last checkbox -- I think we should close this now.

SceneGraph is all-in on Isometry3; we shouldn't try to start shaving that yak under the umbrella of this issue.

mitiguy commented 5 years ago

Closed this issue and opened issue #11549 which deals specifically with SceneGraph.