Fix for describing a Frame using DH Parameters

Hi *!

I am working in a summer school in Italy, and had the good luck of meeting
Ruben Smits. We are using KDL to model the arm of a small humanoid robot
(iCub) and we received the description using DH parameters.

After trying them on KDL, I found out that the DH method in the Frame class
was really using modified DH parameters (as described by John J. Craig in
his book: Introduction to Robotics), but they were referring to them as the
'classic' DH parameters.

To avoid confusion in the future, here is a patch that changes the name of
the previous DH method to mDH, and adds a new DH method that accepts the
'classical' DH parameters.

Note: If you are using the DH method to describe your Frames, you must
switch your code to use mDH instead!

I hope this is useful!

Greetings,

Alexis Maldonado
PhD Student
http://www9.cs.tum.edu/people/maldonad

Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

Fix for describing a Frame using DH Parameters

On Tue, 22 Jul 2008, Alexis Maldonado wrote:

> I am working in a summer school in Italy, and had the good luck of meeting
> Ruben Smits. We are using KDL to model the arm of a small humanoid robot
> (iCub) and we received the description using DH parameters.
>
> After trying them on KDL, I found out that the DH method in the Frame class
> was really using modified DH parameters (as described by John J. Craig in
> his book: Introduction to Robotics), but they were referring to them as the
> 'classic' DH parameters.
>
> To avoid confusion in the future, here is a patch that changes the name of
> the previous DH method to mDH, and adds a new DH method that accepts the
> 'classical' DH parameters.
>
> Note: If you are using the DH method to describe your Frames, you must
> switch your code to use mDH instead!
>
> I hope this is useful!
>
Adjectives such as "new", "classical" or "modified" are meaningless to most
people, especially in a project that wants to serve users in lots of
different communities. So, while I agree with your observation that
multiple versions of DH parameters exist, I suggest that you rather add a
unambiguous adjective to them, such as, for example:
- "DH_Craig1989" if you refer to the definition described in the 1989
version of Craig's book.
- "DH_1955" if you refer to the original ("classical"? :-)) definition as
given in
Denavit, J. and Hartenberg, R. S.
A kinematic notation for lower-pair mechanisms based on matrices
ASME Journal of Applied Mechanics, 23:215-221, 1955.

Herman

Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

Ruben Smits's picture

Fix for describing a Frame using DH Parameters

On Thursday 28 August 2008 13:42:33 Herman Bruyninckx wrote:
> On Tue, 22 Jul 2008, Alexis Maldonado wrote:
> > I am working in a summer school in Italy, and had the good luck of
> > meeting Ruben Smits. We are using KDL to model the arm of a small
> > humanoid robot (iCub) and we received the description using DH
> > parameters.
> >
> > After trying them on KDL, I found out that the DH method in the Frame
> > class was really using modified DH parameters (as described by John J.
> > Craig in his book: Introduction to Robotics), but they were referring to
> > them as the 'classic' DH parameters.
> >
> > To avoid confusion in the future, here is a patch that changes the name
> > of the previous DH method to mDH, and adds a new DH method that accepts
> > the 'classical' DH parameters.
> >
> > Note: If you are using the DH method to describe your Frames, you must
> > switch your code to use mDH instead!
> >
> > I hope this is useful!
>
> Adjectives such as "new", "classical" or "modified" are meaningless to most
> people, especially in a project that wants to serve users in lots of
> different communities. So, while I agree with your observation that
> multiple versions of DH parameters exist, I suggest that you rather add a
> unambiguous adjective to them, such as, for example:
> - "DH_Craig1989" if you refer to the definition described in the 1989
> version of Craig's book.

I agree with this one, DH_Craig1989 is more specific then mDH.

> - "DH_1955" if you refer to the original ("classical"? :-)) definition as
> given in
> Denavit, J. and Hartenberg, R. S.
> A kinematic notation for lower-pair mechanisms based on matrices
> ASME Journal of Applied Mechanics, 23:215-221, 1955.

I would call this just DH because it's an abbreviation of the original authors
names of these parameters and it is widely accepted.

Off course we should put these references in the doxygen information.

Ruben

Fix for describing a Frame using DH Parameters

On Thu, Sep 11, 2008 at 8:41 AM, Ruben Smits
<ruben [dot] smits [..] ...> wrote:
[...]
> I would call this just DH because it's an abbreviation of the original authors
> names of these parameters and it is widely accepted.
>
> Off course we should put these references in the doxygen information.

FWIW, this is also how we did it in BFL: e.g.

k