Can anyone share their DH parameters for a simple PRP manipulator. I have some confusion in setting up the axes and obtaining a solution.
$\begingroup$
$\endgroup$
7
-
1$\begingroup$ The DH parametrisation of a structure has many possible solutions. If you could share the structure and the DH table you already have it would be easier to spot a potential mistake rather then everyone sharing a table which may or may not be usefull to you. $\endgroup$– 50k4Commented Nov 9, 2016 at 7:26
-
$\begingroup$ I'll add a picture. $\endgroup$– SaumitraCommented Nov 9, 2016 at 7:28
-
$\begingroup$ And also the work you have already done (i.e. the current state of the DH table) would help, not to start from zero. $\endgroup$– 50k4Commented Nov 9, 2016 at 7:31
-
$\begingroup$ So...what seems to be the problem? $\endgroup$– 50k4Commented Nov 9, 2016 at 8:27
-
$\begingroup$ The a is thought to be zero and the d assumes values of the offset including the joint variable q. Is this correct? $\endgroup$– SaumitraCommented Nov 9, 2016 at 8:29
|
Show 2 more comments
2 Answers
$\begingroup$
$\endgroup$
In some cases, as are you two cases in the upper part of the image, you might need additional "dummy" coordinate system to end up in the correct orientation. These are just additional lines in the DH table, with most of the coordinates set to 0.
In the upper left solution you might need an additional dummy coordinate system to transform from the 2nd coordinate system to the third one. Just add a dummy coordinate system with a pi/2 z rotation and the x rotation should be pi instead of pi/2 in the transformation which follows (now the 2nd line in the table).
the upper right one needs a dummy transformation with a pi/2 z rotation between system 2 and 3.
The lower left side solution is ok and so is the lower right side one.
$\begingroup$
$\endgroup$
Kinematic frames are different from the physical structure and the working drawings we create can lead us astray. Frames don't necessarily sit on the physical structure. In the frames sketch below frame 1 and 2 overlap. A close mechanical analog with a shift on the x axis between 1 and 2 is shown. However the only requirement is that the z axis match the direction of motion. Also I find it helpful to add my q parameters to the DH table entries so that the offset (the DH constant) is clear.
