Generated IKFast solver for PR2 not respecting joint limits

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Generated IKFast solver for PR2 not respecting joint limits

vchwang
This post was updated on .
I've generated an independent IKFast solver with the latest openrave using the following:

openrave.py --database inversekinematics --robot=robots/pr2-beta-static.zae --manipname=rightarm  --freejoint=r_upper_arm_roll_joint --freeinc=0.01

However, the solutions that come out don't respect joint limits at all. The forward kinematic solution has the end effector at the right place, but many of the joints are completely backwards.

Here's one solution generated:

-2.791392
2.805934
-1.384460
0.988364
-1.544042
1.421673
3.025388

While these are the min joint limits
-0.564602
-0.353600
-0.6500
-2.121300
0
-2.0
0

If I filter out the solutions beyond the joint limits, I don't get any solutions.

Does anyone know why?