Enhance impedance control by nickswalker · Pull Request #79 · TimSchneider42/franky

nickswalker · 2026-03-27T01:13:48Z

Expand the impedance motion stack for both joint-space and Cartesian torque control. Adds tracking/reference-handle based impedance motions, introduces nullspace posture control for Cartesian impedance, adds joint-limit repulsion and shared torque safety utilities, and updates the Python API and docs accordingly.

Add joint impedance motions with joint-limit repulsion.
Add joint impedance tracking via live reference updates.
- This provides a natural solution for trajectory tracking impedance motions (closing Franky and Torque control and Libfranka? #53). This was the primary motivation for this PR.
Refactor the Cartesian impedance API and rename the base class from ImpedanceMotion to CartesianImpedanceBase, now that there is a distinct JointImpedanceMotion class
Add tracking/moving Cartesian references for impedance control.
- This provides a more natural interface for teleoperation uses (closing Smooth 30 Hz joint position control (jerky motion and gripper interference) #76)
Move target_type into fixed-target Cartesian impedance motions.
Add Cartesian nullspace posture control.
Apply the same joint-limit repulsion model to Cartesian impedance.
Expand README/API docs for the impedance motion interfaces.
- At present, the ImpedanceMotion types aren't discussed at all in the readme.

Opening to see if anyone has feedback. I'll add some Python examples and share some videos later in this thread.

TimSchneider42 · 2026-03-27T10:14:25Z

Hi,

I think this PR adds some really important functionality, but I will need some time to review it. I will get back to you once I have taken a look at it. Thank you very much for the PR, though!

Best,
Tim

nickswalker · 2026-03-28T00:55:26Z

Thanks for your willingness to take a look! Here are the two new Python examples running on an FR3.

manual_guidance_joint_impedance.py demonstrates setting joint 6 to be very stiff while leaving the others compliant (a la franka_interactive_controllers), which can sometimes be useful when doing kinesthetic teaching

trim.B82E2E43-2B00-4CC8-8B18-B52AE9AD1165.MOV

cartesian_virtual_fixture.py demonstrates a Python-side loop updating the cartesian reference for the cartesian impedance controller, allowing the user to push the robot along a vertical axis, and resisting all other motion.

trim.247AD8D4-340B-481A-AF0C-A6783A8653C9.MOV

Asunatan · 2026-04-10T11:42:10Z

Hello, thank you for sharing this work—the impedance control shown in the video looks very smooth and promising.

I'm relatively new to robotics and have been evaluating open-source Python libraries for controlling the Franka Emika Panda. Currently, I'm using Deoxys, but I've observed noticeable jitter when performing impedance control, which affects trajectory tracking and force adjustment. I've also looked at FrankaPy, but it doesn't seem to provide a dedicated impedance control interface. Given the performance your library demonstrates, I'm very much looking forward to it…

Brunch-Life · 2026-04-13T08:19:46Z

It's quite a good feature! I'm using it to use gello control franka use joint implance control!

Brunch-Life · 2026-04-14T11:02:41Z

waiting for this to be approved!

nickswalker · 2026-04-14T16:35:04Z

I set the fork to compile and release wheels in the meantime: https://github.com/nickswalker/franky/releases
Also added 0.21.1 libfranka builds, as those are what the videos were recorded with.

TimSchneider42 · 2026-04-15T12:41:12Z

Hey guys,

Sorry for the delay in merging this. I am currently super busy as I am pushing for NeurIPS (deadline is on the 6th of May). I cannot promise to merge this before the deadline, as I would like to review it thoroughly. However, I will try my best to do it as soon as I can.

Best,
Tim

Add JointImpedanceTrackingMotion and CartesianImpedanceTrackingMotion Rename the Cartesian shared base to CartesianImpedanceBase and keep ImpedanceMotion as an alias Extend CartesianImpedanceBase update method with abs_time so it can be passed to whoever is calculating the dynamic reference

Add twist to reference types Without twist, behavior remains as was before With twist, damping only applies to motion that doesn't match the reference twist

Keep target_type on the Cartesian motions that actually interpret a one-time pose target, instead of on the shared impedance base

Way too easy to trip the joint velocity limits by pushing the robot close to a joint limit. Any action the cartesian controller was taking would violate the limit Factor out joint limit repulsion params from joint impedance motion Add saturation to cartesian impedance controller

Use same double-buffered strategy to expose to python Exponentially interpolate gains

An extra backstop against big reference changes causing torque to explode

Negative values could cause the controller to explode Default to reasonable new damping when updating stiffness from Tracker

hugohadfield · 2026-05-08T00:40:49Z

Really looking forward to seeing this land!

The fork's release_wheels workflow produces a wheel with the same file name and version (1.1.3) as the upstream franky-control on PyPI. When users run `pip install --find-links <fork release> franky-control`, pip sees both sources and may pick the upstream PyPI wheel -- which is missing the enhance-impedance-control changes (PR TimSchneider42#79). Patch VERSION in CI to '1.1.3+impedance.libfranka.<ver>' so the produced wheel carries a PEP 440 local-version label that does not exist on PyPI. pip then prefers the local-label wheel from this fork unambiguously. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Signed-off-by: Brunch-Life <brunchlife@hotmail.com>

TimSchneider42 · 2026-05-20T13:26:55Z

Sorry to keep you waiting for so long. I will have some time next week to review the code and merge the PR.

Use std::optional for joint lower/upper limits, and force constraints. Simplifies construction

The fork's release_wheels workflow produces a wheel with the same file name and version (1.1.3) as the upstream franky-control on PyPI. When users run `pip install --find-links <fork release> franky-control`, pip sees both sources and may pick the upstream PyPI wheel -- which is missing the enhance-impedance-control changes (PR TimSchneider42#79). Patch VERSION in CI to '1.1.3+impedance.libfranka.<ver>' so the produced wheel carries a PEP 440 local-version label that does not exist on PyPI. pip then prefers the local-label wheel from this fork unambiguously. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Signed-off-by: Brunch-Life <brunchlife@hotmail.com>

Use composition instead of inheritance for safety param struct Reasonable defaults for per-joint max friction instead of 0s. Actual friction params still default to 0

Pull correct joint bounds based on robot type

nickswalker · 2026-05-25T04:00:14Z

We've been using the JointImpedenceTracker in 0 stiffness for collecting kinesthetic demonstrations lately and noticed that the joints aren't quite transparent. Went ahead and incorporated a simple friction compensation feedforward term (off by default), as I expect this will be a common use case.

nickswalker force-pushed the enhance-impedance-control branch from f0e71ab to e776619 Compare March 28, 2026 01:12

nickswalker marked this pull request as ready for review March 28, 2026 01:13

nickswalker mentioned this pull request Apr 15, 2026

version compatibility #78

Open

nickswalker added 17 commits April 16, 2026 22:58

Add joint impedance motion with joint-limit repulsion

ac713f5

Support moving cartesian references for impedance control

c821755

Add twist to reference types Without twist, behavior remains as was before With twist, damping only applies to motion that doesn't match the reference twist

Move Cartesian target_type into fixed-target motion params

3914539

Keep target_type on the Cartesian motions that actually interpret a one-time pose target, instead of on the shared impedance base

Add nullspace posture term to cartesian impedance controller

2d73a14

Document impedance control motions

e2cd94a

Add nice python wrapper for tracking motions

3c79462

Add support for dynamic impedance gains

6d64899

Use same double-buffered strategy to expose to python Exponentially interpolate gains

Add control rate helper to Python tracker API

e2fce96

Update readme for Tracker API

eea7029

Update Python impedance tracking examples

0ec4a8a

Format code

bbcd0e3

Fix tracker stop with ctrl-C

6d38552

Add task-space error clipping to cartesian impedance controller

073be5c

An extra backstop against big reference changes causing torque to explode

Use default damping for manual guidance demo

8dfad5e

Fix formatting

9672157

nickswalker force-pushed the enhance-impedance-control branch from 55a9773 to f168fd3 Compare April 17, 2026 03:34

Validate stiffness and damping

46852c7

Negative values could cause the controller to explode Default to reasonable new damping when updating stiffness from Tracker

nickswalker force-pushed the enhance-impedance-control branch from f168fd3 to 46852c7 Compare April 17, 2026 03:36

Expose read-only params on impedance motions and trackers

eff8033

Use std::optional for joint lower/upper limits, and force constraints. Simplifies construction

YuqianJiang and others added 5 commits May 24, 2026 23:49

Add optional joint friction compensation

b8b7e51

Move torque param structs into separate header

9042b73

Use composition instead of inheritance for safety param struct Reasonable defaults for per-joint max friction instead of 0s. Actual friction params still default to 0

Consolidate friction demo

35289eb

Update readme for impedance control features

2e26155

Rename impedance tracker examples

5b433be

Pull correct joint bounds based on robot type

nickswalker force-pushed the enhance-impedance-control branch from cdc7abb to 5b433be Compare May 25, 2026 03:55

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Enhance impedance control#79

Enhance impedance control#79
nickswalker wants to merge 24 commits into
TimSchneider42:masterfrom
nickswalker:enhance-impedance-control

nickswalker commented Mar 27, 2026

Uh oh!

TimSchneider42 commented Mar 27, 2026

Uh oh!

nickswalker commented Mar 28, 2026 •

edited

Loading

Uh oh!

Asunatan commented Apr 10, 2026

Uh oh!

Brunch-Life commented Apr 13, 2026

Uh oh!

Brunch-Life commented Apr 14, 2026

Uh oh!

nickswalker commented Apr 14, 2026

Uh oh!

TimSchneider42 commented Apr 15, 2026

Uh oh!

hugohadfield commented May 8, 2026

Uh oh!

TimSchneider42 commented May 20, 2026

Uh oh!

nickswalker commented May 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Conversation

nickswalker commented Mar 27, 2026

Uh oh!

TimSchneider42 commented Mar 27, 2026

Uh oh!

nickswalker commented Mar 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Asunatan commented Apr 10, 2026

Uh oh!

Brunch-Life commented Apr 13, 2026

Uh oh!

Brunch-Life commented Apr 14, 2026

Uh oh!

nickswalker commented Apr 14, 2026

Uh oh!

TimSchneider42 commented Apr 15, 2026

Uh oh!

hugohadfield commented May 8, 2026

Uh oh!

TimSchneider42 commented May 20, 2026

Uh oh!

nickswalker commented May 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

nickswalker commented Mar 28, 2026 •

edited

Loading