Should iCub V2.5 feet size be equal on left/right foot?

[lrapetti]

Assuming that:

• the position of the soles frame inside the foot is mirrored in left and right foot
• the soles reference frames have the same orientation

Then I would expect the dimension of the foot defined on the sole frame: https://github.com/robotology/whole-body-controllers/blob/44ccfa2567938d527406837cfd848775096594bc/controllers/floating-base-balancing-torque-control/app/robots/iCubGazeboV2_5/configStateMachine.m#L155-L156 would be different on right and left foot.

I would expect something like:

left_foot                    = [-0.07  0.12 ;    % xMin, xMax
                                -0.045 0.05 ];   % yMin, yMax  
right_foot                   = [-0.07  0.12 ;    % xMin, xMax
                                -0.05 0.045 ];   % yMin, yMax  

where y dimensions are flipped on one foot.

@gabrielenava are my assumptions wrong? Am I missing something?

[gabrielenava]

I think you are right @lrapetti, I suspect the feet size was never corrected as the asymmetry in the y direction seems small and the yoga motion in particular was always conservative enough to not be affected by this error. Could you open a PR to fix the feet size?

[lrapetti]

Could you open a PR to fix the feet size?

Sure, I can handle it. Since the same variable feet_size is used both on left and right foot, we can keep the change minimum simply by using a symmetric value like:

 feet_size                    = [-0.07   0.12 ;    % xMin, xMax 
                                 -0.045 0.045 ];   % yMin, yMax 

what do you think?

[gabrielenava]

Sure, I can handle it. Since the same variable feet_size is used both on left and right foot, we can keep the change minimum simply by using a symmetric value like:

feet_size = [-0.07 0.12 ; % xMin, xMax -0.045 0.045 ]; % yMin, yMax what do you think?

fine to me :-)

[lrapetti]

PR opened https://github.com/robotology/whole-body-controllers/pull/137

[gabrielenava]

PR merged and closed, thanks @lrapetti! closing the issue.