search

isri-aist / BaselineWalkingController

Humanoid walking controller with various baseline methods
https://isri-aist.github.io/BaselineWalkingController
BSD 2-Clause "Simplified" License
110 stars 19 forks source link

Robot falls over in the CI simulation tests #18

Closed mmurooka closed 11 months ago

mmurooka commented 11 months ago

https://github.com/isri-aist/BaselineWalkingController/actions/runs/6133956095

The robot falls over in CI simulation tests of some conditions. It did not fail last week in scheduled CI runs, but started failing this week. There have been no changes to BWC in the meantime, and no major changes to mc_rtc (mc_rtc is installed from the head repository, not stable).

As noted in the comment at https://github.com/isri-aist/CnoidRosUtils/commit/dc57984656536b575a50b65003dbb4edc8587129 is the version of Choreonoid in apt recently updated? (I just want to know for reference) @gergondet

gergondet commented 11 months ago

Hi @mmurooka

I think the issue might be related to https://github.com/jrl-umi3218/mc_rtc/pull/388

It seems what is computed in the state-observation library and the implementation and the old implementation in mc_rtc are in fact different.

I'm looking into it a little more.

mmurooka commented 11 months ago

Thank you. Following the last issue of convex transformation, BWC's scheduled CI detected mc_rtc regression again, so maybe it would be good to set up a mechanism to call BWC's CI every time mc_rtc main branch is pushed

mmurooka commented 11 months ago

I will cloes this issue now that the CI has passed.