From: Geoffrey B. <g....@au...> - 2006-02-23 22:39:56
|
The mapping would be handled in the driver, so that should be easy enough to set up how you want. The actarray can treat the gripper like another actuator if that's how the driver is written. Currently the p2os driver provides an array of 6 actuators for the 5DOF+gripper pioneer arm. There is a gripper interface, but it's highly specific to the pioneer gripper (not the one on the end of the arm). I think it's had a todo on it to make it more generic for a while. I'd quite like to sort that out at some point, but I need to know what a gripper interface needs (it might even be easiest to just use actarray for generic grippers), and I think we should get the player 2 release out of the way first. Until then treating it as either another actuator on the actarray or using a separate position1d/actarray device for it is the best option I can think of. Geoff Reed Hedges wrote: > > actarray would be the right one, I think, then it would work for either > one PowerCube pan/tilt, or for a whole series of single joints with a > PowerCube at the base and/or the end (we call that the Power Arm or > something like that). You would just have to get the mapping right > between the actarray joint identifiers and whatever id's the two joints > in the powercube happen to be using. For example, our in-house can > powercube has identifiers 5 and 6 but it could vary. > > Does actarray treat a gripper-style end effector the same as other > joints? Or would we create a seperate driver for that which implements > the "gripper" interface (IIRC the Amtec gripper is just another joint > but interprets positions differently). > > (Also, when I said "two position2d devices" before I meant "two > position1d devices" (ie motor devices).) > > Reed -- Robotics research group, University of Auckland http://www.ece.auckland.ac.nz/~gbig005/ |