1
$\begingroup$

I am implementing a Denavit-Hartenberg forward transform for a 3-axes CNC mill. I know that the kinematic for such a machine is trivial and doesn't need DH, but I need to make appliable for other robots too. My implementation does the math correctly(I've verified that with another tool), but the transformation doesn't give me the results I would expect.

I assume that for 3-axes cartesian robot with orthogonal prismatic joints(=CNC mill) the resulting transformation matrix should give me the input parameters(d1-d3) back in its translation vector, but it somehow doesn't. Also, the resulting orientation matrix should have some "nice" values(90, 180, 270, etc.) and no odd ones(0.0528, 0.5987, etc,.). Is my assumption wrong?

$\endgroup$

1 Answer 1

1
$\begingroup$

Okay, I feel really dumb now. Turns out my sine/cosine functions take radiant as input rather than degrees. As expected, the resulting transform matrix has "nice" values. Since my implementation and the other tool, which I checked the results with, both use rad as input. Consequentially I assumed the math was correct, which was false and resulted in odd numbers.

$\endgroup$
1
  • 1
    $\begingroup$ Sounds like a unit test could have saved you a lot of time :) $\endgroup$
    – user9092
    May 3, 2015 at 16:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.