ergoCub 1.1 S/N:001 – Problems when loading configuration files for torque control parameters

[DanielePucci]

Robot Name 🤖

ergoCub 1.1 S/N:001

Request/Failure description

Today, we started with @isorrentino to work to test the Yoga demo on the real robot after the preliminary results in https://github.com/ami-iit/element_qp-reactive-control/issues/70.

So, we started modifying the files for loading the parameters of torque control. See the commit ami-iit/robots-configuration@f6fe83c

Unfortunately, these parameters seem not loaded in the motor gui

And we cannot at the moment test the torque control since adding them manually for all the joints is a very long process

Detailed context

This is the output of yarprobotinterface
output_interface.txt

@Nicogene @sgiraz @maggia80

Additional context

No response

How does it affect you?

No response

[sgiraz]

Hi @DanielePucci,

The Parser seems to parse correctly the values set in config files:

...

(stictionUp 0 0 0) (stictionDown 0 0 0) (kff 1 1 1) (viscousPos 0 0 0) (viscousNeg 0 0 0) (coulombPos 0 0 0) (coulombNeg 0 0 0) (velocityThres 0.1 0.1 0.1) (filterType 0 0 0) (ktau 176 176 -176)"), ("OTHER_CONTROL_PARAMETERS" [group] = "(deadZone 0.0049 0.0049 0.0049)")]

...

I'm not sure, but as a first quick view to the info provided, it could be a (wrong) cast from float to int that trunkates the 0.1 to 0.
By the way, we'll perform some tests in during our next sprint (starting from Monday 22).

cc @Nicogene

[DanielePucci]

Hi @sgiraz, thanks for the hint. We tried to set all parameters as integers, but it does not seem to solve the issue

Here's the log log_ergocub-torso_yarprobotinterface_18098.txt

[sgiraz]

Hi @DanielePucci,

I fixed a bug in icub-main that was preventing the parsing of torque parameters when the (⚠️ deprecated) kbemf parameter was missing. Here's the PR that addresses the issue:

• Bugfix kbemf set to 0 by default when not found icub-main#935

cc @isorrentino @GiulioRomualdi @AntonioConsilvio

[DanielePucci]

Thanks a lot @sgiraz, hopefully tomorrow we are going to give it a try!

[isorrentino]

Today, @DanielePucci and I tested the PR. Now everything works fine. Thanks a lot.

[sgiraz]

Hi @isorrentino, @DanielePucci

Thank you for your precious feedback.

I proceed closing this issue! ✅

Cc @Nicogene @AntonioConsilvio