Notes taken by Andy Zelenak
Attendees
- Dave Coleman
- Jon Binney
- Robert Haschke
- Ian McMahon
- Bryce Willey
- Gijs vd. Hoorn
- Jorge Nicho
- Nathan Brooks
- Mike Lautman
- Henning Kaiser
- Andy Zelenak
- Yu Yan
The “!” is officially gone from the title of MoveIt
Jog Arm PR has been merged
- real-time, Jacobian-based jogging at ~500-1000 Hz control rates
- Has been tested mostly on UR robots
- Will not work yet for most industrial robots because it’s not jerk-limited
- Exists in moveit_experimental for now
Discussion of a new MoveIt Advanced Tutorials webpage
- The MoveIt tutorials page is getting busy
- May create a new page for larger demos/ integration tests/ tutorials
- Yu to provide an example he has already created
OMPL Planning Plugin
- Continuing the efforts of Zak Kingston and Ryan Luna to make the OMPL planning plugin more modular, i.e. you’ll be able to write your own sampling-based planner (or any other type of planner) in OMPL and use it in MoveIt without compiling MoveIt from source.
- Starting off by splitting up https://github.com/ros-planning/moveit/pull/897
- First PR in the series: https://github.com/ros-planning/moveit/pull/1428
GSoC Progress (Bryce Willey and Felix von Drigalski)
- MoveIt requested 1-2 slots through OSRF for students to work on the proposed projects focusing on collision detection and TrajOpt integration (https://github.com/osrf/osrf_wiki/wiki/GSoC#moveit-projects).
- OSRF received many of it’s requested spots for Google Summer of Code students.
- The selected students will be announced May 6th, so be on the look out for new faces in the community.
Bryce was voted in as a full MoveIt maintainer with write access
A workaround for testing MoveIt packages outside of Travis CI
- Travis CI limits test time to 50 minutes
- The MoveIt repo is large enough to often exceed that limit
- Mike is working on a Docker method to run tests locally
- Robert recommends a different approach, will work with Mike offline
How do the maintainers like the new auto-ping feature
- General approval
- Mike will continue providing a monthly report
ROS2 progress
- Accutronic is making good progress
- It’s uncertain if the Dashing release will be met
- Emphasize on high-quality, well-tested code vs. Dashing release, if needed
- This initial effort is mostly a port, an extensive rewrite is desired for later
Potential ROS2 features
- Planning can take several seconds, during which the PlanningScene could change
- Would be nice to update the PlanningScene faster in ROS2