In order to improve the time taken for IK computations, I decided to implement the IKfast solver on my 6 DoF robot arm in Moveit. Everything appears to be working fine except for one joint that doesn't move, this specific joint is shown in the attached image URL.
For example, if I move the arm to the left/right with a constant orientation, the pan joint will rotate but this other "yaw joint" remains stationary so the end effector doesn't compensate its orientation.
I have confirmed this in both the rviz planner, and in my own CPP code - where the IK output is 0.
http://tinypic.com/r/2nvypvs/9
↧