Re: Offset in eyes' tilt - icub at Lisbon

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: Offset in eyes' tilt - icub at Lisbon

Plinio Moreno-2
Dear all,

A follow up on my previous post,

On 10/28/2014 05:19 PM, Plinio Moreno wrote:
> Dear Stefano and all,
>
> we realized that the home position of the eyes's tilt is shifted about
> 7 degrees. Since now we are running the robotInterface, we want to
> know where should we change the home position of the tilt to match the
> description of the kinematic chain, or if we need to rerun the head
> calibration.

More information about this offset:

When the eyes' tilt is at zero degrees, the robot should be pointing
towards the horizon according to the Denavit-Hartenberg reference.
However, this is not happening now (the robot is looking a tiny bit
down). This is causing the control gaze and the stereo to have an offset
of about 12 cm.

Best regards,

Plinio

>
> Best regards,
>


--
___________________

Plinio Moreno López
http://www.isr.ist.utl.pt/~plinio
[hidden email]
___________________


------------------------------------------------------------------------------
_______________________________________________
Robotcub-hackers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/robotcub-hackers
Reply | Threaded
Open this post in threaded view
|

Re: Offset in eyes' tilt - icub at Lisbon

pattacini
Hi Plinio,

You should compute the "Delta" values according to the procedure described here: http://wiki.icub.org/wiki/HeadFineCalibration.
Then, put those "Delta" in the configuration file read by the robotInterface: iCubLisboa01\hardware\motorControl\icub_head.xml.

Given the file branching, you should update the file in repository issuing a PR via github and then also the local file on your system.

However, bear in mind that aligning the robot's pupils does not mean that the camera CCD's are aligned with the CAD model: it does certainly help out but is not decisive.

Ciao
Ugo

> -----Original Message-----
> From: Plinio Moreno [mailto:[hidden email]]
> Sent: 28 October 2014 18:29
> To: [hidden email]
> Subject: Re: [rc-hackers] Offset in eyes' tilt - icub at Lisbon
>
> Dear all,
>
> A follow up on my previous post,
>
> On 10/28/2014 05:19 PM, Plinio Moreno wrote:
> > Dear Stefano and all,
> >
> > we realized that the home position of the eyes's tilt is shifted about
> > 7 degrees. Since now we are running the robotInterface, we want to
> > know where should we change the home position of the tilt to match the
> > description of the kinematic chain, or if we need to rerun the head
> > calibration.
>
> More information about this offset:
>
> When the eyes' tilt is at zero degrees, the robot should be pointing towards
> the horizon according to the Denavit-Hartenberg reference.
> However, this is not happening now (the robot is looking a tiny bit down).
> This is causing the control gaze and the stereo to have an offset of about 12
> cm.
>
> Best regards,
>
> Plinio
>
> >
> > Best regards,
> >
>
>
> --
> ___________________
>
> Plinio Moreno López
> http://www.isr.ist.utl.pt/~plinio
> [hidden email]
> ___________________
>
>
> -----------------------------------------------------------------------------
> -
> _______________________________________________
> Robotcub-hackers mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/robotcub-hackers

------------------------------------------------------------------------------
_______________________________________________
Robotcub-hackers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/robotcub-hackers
Reply | Threaded
Open this post in threaded view
|

Re: Offset in eyes' tilt - icub at Lisbon

Plinio Moreno-2
Hi Ugo,

thanks for the pointer to that. Now the eyes' tilt matches the
Denavit-Hartenberg zeros.

Cheers,

Plinio

On 10/28/2014 06:54 PM, Ugo Pattacini wrote:

> Hi Plinio,
>
> You should compute the "Delta" values according to the procedure described here: http://wiki.icub.org/wiki/HeadFineCalibration.
> Then, put those "Delta" in the configuration file read by the robotInterface: iCubLisboa01\hardware\motorControl\icub_head.xml.
>
> Given the file branching, you should update the file in repository issuing a PR via github and then also the local file on your system.
>
> However, bear in mind that aligning the robot's pupils does not mean that the camera CCD's are aligned with the CAD model: it does certainly help out but is not decisive.
>
> Ciao
> Ugo
>
>> -----Original Message-----
>> From: Plinio Moreno [mailto:[hidden email]]
>> Sent: 28 October 2014 18:29
>> To: [hidden email]
>> Subject: Re: [rc-hackers] Offset in eyes' tilt - icub at Lisbon
>>
>> Dear all,
>>
>> A follow up on my previous post,
>>
>> On 10/28/2014 05:19 PM, Plinio Moreno wrote:
>>> Dear Stefano and all,
>>>
>>> we realized that the home position of the eyes's tilt is shifted about
>>> 7 degrees. Since now we are running the robotInterface, we want to
>>> know where should we change the home position of the tilt to match the
>>> description of the kinematic chain, or if we need to rerun the head
>>> calibration.
>> More information about this offset:
>>
>> When the eyes' tilt is at zero degrees, the robot should be pointing towards
>> the horizon according to the Denavit-Hartenberg reference.
>> However, this is not happening now (the robot is looking a tiny bit down).
>> This is causing the control gaze and the stereo to have an offset of about 12
>> cm.
>>
>> Best regards,
>>
>> Plinio
>>
>>> Best regards,
>>>
>>
>> --
>> ___________________
>>
>> Plinio Moreno López
>> http://www.isr.ist.utl.pt/~plinio
>> [hidden email]
>> ___________________
>>
>>
>> -----------------------------------------------------------------------------
>> -
>> _______________________________________________
>> Robotcub-hackers mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/robotcub-hackers


--
___________________

Plinio Moreno López
http://www.isr.ist.utl.pt/~plinio
[hidden email]
___________________


------------------------------------------------------------------------------
_______________________________________________
Robotcub-hackers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/robotcub-hackers