Q: How to obtain position-only inverse kinematics w/ trac_ik?

[Joseph Coombe]

Hello, I'm using trac_ik_python for a robot arm.

In certain cases, I'm only interested in achieving a certain end-effector position (do not care what its orientation is).

I've tried set really large orientation bounds - and it has worked for all cases I've tested so far - but I'm not sure if this is the best way.

target_jt_ang = ik_solver.get_ik(seed_state,
    0.45, 0.1, 0.3,  # X, Y, Z end-effector target pos
    0.0, 0.0, 0.0, 1.0,  # rx, ry, rz, rw end-effector orientation < don't care
    0.01, 0.01, 0.01,  # X, Y, Z bounds
    1000, 1000, 1000  # Rotation X, Y, Z bounds
    )

Thank you

[Patrick Beeson]

That is the best way. Are there are cases where is doesn't work? It should always work, but I haven't done much testing on the swig wrapper that was written by another individual.

[Joseph Coombe]

Awesome, I'm still testing to get better insight as to why the solver fails in some cases - not sure if it is related to position-only IK. Will open new issue if I get more information.

Also, thank you very much for writing this. It is blazingly fast!

[Joseph Coombe]

Quick update: I'm a moron. The case where the solver was failing, it was because I was trying to move a position that was physically impossible to achieve with safety limits I had placed on the motors.

Did not realize this until I tried to manually move the arm through the trajectory myself.