Skip to content

Bugfix/fix capture region and rate limit #1110

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 2 commits into from
Nov 18, 2025
Merged

Bugfix/fix capture region and rate limit #1110

merged 2 commits into from
Nov 18, 2025
+42 −6

Conversation

@rjgriffin42
Copy link
Member

@rjgriffin42 rjgriffin42 commented Nov 18, 2025

No description provided.

rjgriffin42
rjgriffin42 commented Nov 18, 2025
View reviewed changes
...monWalkingControlModules/capturePoint/stepAdjustment/ErrorBasedStepAdjustmentController.java
else
{
computeBestReachabilityConstraintToUseWhenIntersecting();
getSelectedRegion().orthogonalProjection(adjustedSolution);
Copy link
Member Author

@rjgriffin42 rjgriffin42 Nov 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I switched this back to what it was previously before the previous change. As written in the code, if returning the selected region, it returned the intersection between the reachable region and the capture point. Previously, the conditions would return this if the intersections were valid, so that there was an intersection, and if there were no intersections, it would return just the reachable region. The change in the PR returned only the selected one, so that the projection failed.

I changed this, so that if there's an intersection, it returns the selected reachable region, while refactoring the previous method to say "selectedReachableCaptureRegion", which is what is returned if these regions are valid.

rjgriffin42
rjgriffin42 commented Nov 18, 2025
View reviewed changes
...src/main/java/us/ihmc/commonWalkingControlModules/capturePoint/controller/ICPController.java

solver.resetCoPFeedbackConditions();
solver.resetFeedbackDirection();
solver.resetFeedbackConditions();
Copy link
Member Author

@rjgriffin42 rjgriffin42 Nov 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I consolidated this logic into a single reset method, while also removing the limits.

rjgriffin42
rjgriffin42 commented Nov 18, 2025
View reviewed changes
...src/main/java/us/ihmc/commonWalkingControlModules/capturePoint/controller/ICPController.java
{
solver.removeMaximumFeedbackMagnitude();
solver.removeFeedbackRateLimit();
}
Copy link
Member Author

@rjgriffin42 rjgriffin42 Nov 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Previously, what happened is that we wouldn't add the magnitude limits or rate limits if it was in a reset. However, those values were never getting reset inside the ICPControllerQPSolver. So it was applying whatever they were the last time they were set. So we were never actually removing these constraints.

rjgriffin42
rjgriffin42 commented Nov 18, 2025
View reviewed changes
.../java/us/ihmc/commonWalkingControlModules/capturePoint/controller/ICPControllerQPSolver.java
resetCMPFeedbackConditions();
resetFeedbackDirection();
removeFeedbackRateLimit();
removeMaximumFeedbackMagnitude();
Copy link
Member Author

@rjgriffin42 rjgriffin42 Nov 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we were never calling "resetCMPFeedbackConditions". While this didn't cause any problems, it is a bug. The other two are the things that reset those rate limits and magnitude limits each time so they can be selectively enforced.

@PotatoPeeler3000 PotatoPeeler3000 merged commit 0eb9c5e into demo/alex-nov-2025 Nov 18, 2025
60 of 67 checks passed
@PotatoPeeler3000 PotatoPeeler3000 deleted the bugfix/fix_capture_region_and_rate_limit branch November 18, 2025 21:44
@PotatoPeeler3000 PotatoPeeler3000 mentioned this pull request Nov 21, 2025
Merged
PotatoPeeler3000 added a commit that referenced this pull request Nov 21, 2025
@PotatoPeeler3000 @LuigiPenco93
@reesep5 @calvertdw @rjgriffin42 @TomaszTB @ihmc-rosie @ds58
Outside demo branch with Alexander on the week before Thanksgiving in…
99ba635 
… Novdember 2025 (#1121)

* alex teleop

* ghost robot

* fixed a weird constructor in OpenAlexanderRobotModel

* tracker roles

* disable kst server

* Got local logging working and tuned the realsense and height map for real world testing (#1063)

* Settup logging to work on both the high and low level process, fixed a couple of bugs in the starting, and naming of the thing

* This shouldn't be a hard coded value, neesds to be the cell size from the parameters

* Shrink height map to reduce networking lag

* Adjust realsense transform, adjust local logging for testing

* Tested more of the logging onboard. Update is either the server, or the intraprocess logger, not both. Optimized the IntraprocessYoVariableLogger so now the longest part is compressing the data before it gets written to disk

* Fix gradle deploy, remove link in warning prints

* Fix VR finger streaming.

---------

Co-authored-by: Alexander OCU <rosie@ihmc.org>
Co-authored-by: VR-OCU <lpenco@ihmc.org>

* Setup behavior tree scene. (#1065)

* added a getter (#1069)

* Fix sync feedback loop bug with footstep plan action interactables. (#1070)

* minor null check in multi thread factory

* undo robot model stuff

* remove artifcats from merge

* remove artifcats

* tuned kst

* fixes pr

* Revert "remove artifcats from merge"

This reverts commit ca073ed.

* fix with develop

* Update OpenAlexanderMomentumOptimizationSettings.java

* Bugfix/fix pelvis rotation on delays (#1106)

* started working on a fix for pelvis rotationd elays

* rolled back on how the time in the phase is found, and started adding a decay factor so the feedforward values dont go to the moon

* fixed a bound inclusiveness problem

* added some changes to prefer previous reachable regions to yield more… (#1105)

* added some changes to prefer previous reachable regions to yield more consistent, smoother results

* reverted line change

---------

Co-authored-by: Nick Kitchel <59104880+PotatoPeeler3000@users.noreply.github.com>

* remove yaw component from angular velocity filtering of the foot (#1107)

Co-authored-by: Nick Kitchel <59104880+PotatoPeeler3000@users.noreply.github.com>

* add a filter for big downward forces on the second threshold (#1108)

Co-authored-by: Nick Kitchel <59104880+PotatoPeeler3000@users.noreply.github.com>

* [Feature] CSG Heartbeat (#1104)

* Make ROS2HeartbeatMonitor garbage free

* Add heartbeats to CSG objects

* Added destroy method to CSGROS2CommunicationHelper and cleaned up usage in the panel

---------

Co-authored-by: Alexander OCU <rosie@ihmc.org>
Co-authored-by: Nick Kitchel <59104880+PotatoPeeler3000@users.noreply.github.com>

* Add null check to ImageSensor

* Bugfix/fix capture region and rate limit (#1110)

* fixed the reachability constraint return

* properly reset the feedback conditions for the ICP feedback

* Feature/rdx buttons (#1111)

* set up the buttons for whether the steps are adjustable

* added a swing height slider to rdx

* Remaking the demopose branch for some conflict reason (#1118)

Co-authored-by: Alexander OCU <rosie@ihmc.org>

* Intraprocess logger safety (#1114)

* Fix Ability hand VR rotator streaming.

* Catch exception in intraprocess writer compression thread and also check that the directory is usable

* Minor

* Minor

* Revert thing

* Switch CountDownLatch to object wait/notify

---------

Co-authored-by: Duncan Calvert <dcalvert@ihmc.us>
Co-authored-by: ihmc-rosie-windows <rosie@ihmc.org>

* Fix the infinite loop with the height map in the controlller, cleaned up the UI for RDX with CSG, and removed a duplicate updater.  (#1116)

* Fix Ability hand VR rotator streaming.

* Fix the bugs with the CSG and the height map, and more

* made csg rdx panel buttons for snapping to heightmap and accounting for ground height drift/footstep z error

---------

Co-authored-by: Duncan Calvert <dcalvert@ihmc.us>
Co-authored-by: ihmc-rosie <rosie@ihmc.us>
Co-authored-by: Alexander OCU <rosie@ihmc.org>

* upped the icp distance threshold for falling

* Feature/height map noise hacking (#1119)

* Add option to parameters to tune the minimum acceptable depth values

* Tuned value

* Update transform and parameter for swing

---------

Co-authored-by: ihmc-rosie-windows <rosie@ihmc.org>

* Added an option to log a full resolution heartbeat for the depth image (#1120)

Co-authored-by: nkitchel <nkitchel@ihmc.org>

* Revert "Added an option to log a full resolution heartbeat for the depth image (#1120)"

This reverts commit 2cdb895.

* Add destroy method to the AvatarStepGeneratorThread

---------

Co-authored-by: Luigi Penco <lpenco@ihmc.org>
Co-authored-by: Alexander OCU <rpeterson@ihmc.org>
Co-authored-by: ihmc-rosie-windows <rosie@ihmc.org>
Co-authored-by: Duncan Calvert <dcalvert@ihmc.org>
Co-authored-by: Robert Griffin <rgriffin@ihmc.us>
Co-authored-by: LuigiPenco <161648875+LuigiPenco93@users.noreply.github.com>
Co-authored-by: Tomasz Bialek <103042423+TomaszTB@users.noreply.github.com>
Co-authored-by: ihmc-rosie <rosie@ihmc.us>
Co-authored-by: ds58 <30220598+ds58@users.noreply.github.com>
Co-authored-by: Duncan Calvert <dcalvert@ihmc.us>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@Beomyeong Beomyeong Beomyeong approved these changes

@reesep5 reesep5 reesep5 approved these changes

@james-p-foster james-p-foster Awaiting requested review from james-p-foster

@PotatoPeeler3000 PotatoPeeler3000 Awaiting requested review from PotatoPeeler3000

@stefanfasano stefanfasano Awaiting requested review from stefanfasano

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@rjgriffin42 @Beomyeong @reesep5 @PotatoPeeler3000