Moveit fails on multistep execution
I have a basic ros (melodic) python node where i can enter multiple points and the robot will make a plan to go to each point in order, using the last end state as the new start state. Nothing really complicated here and when i run it in the simulator, all works well, the fake robot executes as expected. However when i execute this on the real robot, it never makes it past the first step throwing the error that the start state of the plan is different than the current state. This makes any movement node I write go very slow because for each movement the robot has to plan one movement, move, then repeat. Is there a way to either allow some tolerance on the start state? Or should i perhaps be tightening my end state tolerance?
Edit: Tested on yaskawa motoman gp12 (6 dof)
Is this with your Motoman setup?
That would be important information to mention, as there are various components which could be complaining here.
@gvdhoorn yes it is, my bad i forgot to include it. Going to try it on a fanuc lr mate later this month to see if the issue persists but i figured i would see if this is normal.
I would recommend you post an example of the error message you mention.
From your description, it could be both MoveIt and/or
motoman_driverwhich is complaining.Also: please describe what you're trying to achieve a bit. What sort of process are you programming these robot motions for? Should they really be point-to-point motions or should it actually be one continuous trajectory?