MoveIt Maintainer Meeting - Aug 25

The next Manipulation WG Meeting / MoveIt Maintainer Meeting will be 2022-08-25T15:00:00Z.


  • Old Business
  • MoveIt 2 Update
  • Support for HPP-FCL as collision checker (Vatan)
  • Discussion about collision libraries in general (Henning)
    • How is current benchmarked performance actually compairing to tesseract
    • What are the current shortcomings? Bullet incomplete, CCD,…


Real-time meetings by Google. Using your browser, share your video, desktop, and presentations with teammates and customers.

Phone Numbers (US)

+1 347-486-5750

PIN: 644 991 049#

I’m afraid I miss a lot of these meetings this year… :slightly_frowning_face:
Here are some thoughts from my side.

  • Support for HPP-FCL as collision checker (Vatan)

Great to see my suggestion being picked up!

  • How is current benchmarked performance actually compairing to tesseract

As Levi pointed out multiple times MoveIt’s use of FCL is rather convoluted, partially for historical reasons, partially by (current) design. The caches can help a lot in some cases, but not in general.
I believe a big point of the “by design” part is the representation of the robot as a separate structure which needs to be recreated for each check.

  • What are the current shortcomings? Bullet incomplete, CCD,…

A user invested a lot of time last year in improving our bullet CCD, but the work got stuck with some suggested major improvements: Interpolate paths for less under-approximation in ccd of bullet checker (#2889) by werner291 · Pull Request #2897 · ros-planning/moveit · GitHub

From my perspective by far the biggest shortcoming in URDF/MoveIt though is the inability to mark meshes as convex.
With convex mesh support many realistic collision checks could speed up tremendously (though models would actually need to be changed to ensure convex meshes and mark them as such).


Here are my notes from the meeting:

  • MoveIt release for bug-fixes in Humble
  • [Vatan] HPP FCL
    • Fork of FCL that diverged significantly
    • Has been in legged robotics applications
    • Saw Michael’s issue and has spent a few days integrating it
    • Benchmarks showing it is ~2x faster than FCL
    • Next steps: Fixing linker errors and benchmarking
  • [Simon] Fcl combined collision check by simonschmeisser · Pull Request #3167 · ros-planning/moveit · GitHub
    • PR to improve performance of collision checking in some cases
  • [Vatan] According to tessaract’s plots FCL is 2x faster than bullet
  • [Henning] I think continuous collision checking could be realy useful
    • Only consider motion between two states and only check the spanned motion between the two states
    • Do not need to discritize a linear interpolation between states and collision check each point
    • Some interfaces already exist in collision checkers for this and we could implement usign it
    • [Vatan] HPP FCL doesn’t do CCD anymore.
    • [Andy] Could really speed up planning but I don’t think we need it
    • [Henning] I think multi-trajectory execution could also proffit from CCD
  • [Vatan] Do we have a benchmark of which parts of moveit takes the most compute when planning
    • [Henning] MoveIt benchmark gives you benchmarks of planning
    • [Simon] GitHub - KDAB/hotspot: The Linux perf GUI for performance analysis.
    • [Henning] It is easy to observe that collision checking is the bottleneck by changes in your environment affects planning time
    • [Vatan] I agree, but do we have data
    • [Simon] You can also just do less collision checks if you can reason at a higher level to find solutions you don’t have to collision check. We should find ways to not use RRTConnect and find better solutions.
    • [Henning] Maybe the most performance improvement could be in higher level reasoning. Rejection sampling is really expensive because you do all the work only to discard the result.
    • [Tyler] I’ve had good results with Intel® VTune™ Profiler when benchmarking small bits of code
  • [Henning] What is the future of moveit benchmark
    • [Simon] CaptianYoshi are doing to work on a paper using moveit1
  • [Andy] MoveIt 2 tutorials are broken


  • Robert Haschke
  • David Lu!!
  • Vatan Aksoy Tezer
  • Sebastian Jahr
  • Nathan Brooks
  • Michael Ferguson
  • Henning Kayser
  • Cassidy Elliott
  • Cory Crean
  • Paul Gesel
  • Eric Holum
  • Jafar Uric
  • Andy Zelenak
  • Simon Schmeisser
  • Marq Rasmussen
  • Tyler Weaver
1 Like